Physiological perfusion of human vasculature reveals a YAP/TAZ-Apelin switch linking intraluminal flow to endothelial state transitions and vessel remodeling

The authors developed VIVOS, a perfused 3D human vascular platform that reveals how physiological intraluminal flow drives endothelial state transitions and vessel remodeling through a YAP/TAZ-Apelin signaling switch, while also enabling the modeling of flow-regulated pathologies like arteriovenous malformations.

The Gist

Imagine trying to build a tiny, living city inside a petri dish to test new medicines. The problem? In a real city, people and goods move around via roads, highways, and rivers. But in most lab-grown tissues, nutrients and signals have to "walk" through the air or gel to get where they need to go. This is like trying to feed a whole city by dropping a single sandwich from a helicopter every day—it works for a small village, but for a big city, people in the middle would starve.

This paper introduces a revolutionary new system called VIVOS (Vascularized In Vitro Organ Systems) that solves this problem by building a fully human "circulatory system" right on the lab bench.

Here is the story of what they discovered, broken down into simple concepts:

1. The "Heart" of the Machine: VIVOS

The researchers built a device that acts like a tiny, artificial heart. Instead of letting nutrients drift slowly, they use a special pump (an impeller, similar to a boat propeller) to push fluid through millimeter-sized blood vessels they grew in a lab.

• The Analogy: Think of traditional lab dishes as a still pond where you have to wait for a leaf to drift to the other side. VIVOS is like a rushing river. It forces water (and the nutrients inside it) to flow continuously through the tissue, just like blood flows through your veins.
• The Result: They can now grow large, complex human tissues (like mini-brains, lungs, and retinas) and keep them alive and healthy because the "river" delivers food exactly where it's needed.

2. The "Traffic Light" for Cells: Flow Changes Behavior

The most exciting discovery is that the speed of the water flowing through these vessels actually changes the personality of the cells lining the walls (the endothelial cells).

• The Analogy: Imagine a busy highway. When traffic is moving fast and smoothly (laminar flow), the road workers (cells) calm down and start building sturdy, straight roads. When traffic is slow or chaotic, the workers get stressed and start frantically building new, messy side-roads (sprouting).
• The Science: They found a molecular "switch" inside the cells involving two proteins called YAP/TAZ.
  • Fast Flow (Healthy): The water pressure tells YAP/TAZ to "shut up." This turns off the "sprouting" signal and turns on a signal called Apelin. The cells decide, "We are good here, let's stabilize the vessel."
  • Slow Flow (Unhealthy): YAP/TAZ stays active. This triggers the cells to become "tip cells" (like construction crews) that try to sprout new vessels everywhere, leading to messy, tangled networks.

3. The "Apelin Switch": A Molecular Conversation

The researchers discovered that the cells talk to each other using a chemical messenger called Apelin.

• The Analogy: Think of the blood vessel as a construction site.
  • The Tip Cells (the scouts) shout, "We need more roads!" (They produce Apelin).
  • The Stalk Cells (the builders) listen and say, "Okay, we'll build here," but only if they hear the right signal.
  • The Twist: When the water flows fast, the "scouts" stop shouting. The "builders" hear the silence, realize the road is stable, and stop building. It's a self-regulating system that prevents the vessel from growing out of control.

4. Modeling Disease: The "Broken Dam" (HHT)

The team used VIVOS to model a rare genetic disease called Hereditary Hemorrhagic Telangiectasia (HHT). In this disease, patients have "short circuits" in their blood vessels (arteriovenous malformations) where blood rushes through without slowing down, causing dangerous bleeds.

• The Experiment: They created vessels with a broken "brake system" (missing a protein called Endoglin).
• The Observation: Just like in the disease, the vessels became huge, dilated, and the blood rushed through them too fast.
• The Cure Test: They added a drug called BMP9. In the VIVOS system, BMP9 acted like a "traffic cop." It told the chaotic, fast-flowing vessels to shrink back down and stop the reckless growth. This proves that VIVOS can be used to test drugs that might fix these broken vessels in real humans.

Why This Matters

Before VIVOS, scientists had to use animals to study how blood flow affects disease, or they had to use flat layers of cells that didn't behave like real 3D tissues.

• The Big Picture: VIVOS is a fully human, animal-free test drive for vascular diseases. It allows scientists to see exactly how blood flow changes cell behavior, how diseases form, and whether new drugs can fix the problem, all in a controlled, human environment.

In short: They built a tiny, flowing river in a dish to teach human cells how to behave like they do in a real body, discovered the "traffic light" that controls their growth, and used it to test a cure for a dangerous blood vessel disease.

Technical Summary

1. The Problem

Current preclinical models for studying vascular biology, such as organoids and organ chips, suffer from two major limitations:

• Lack of Physiological Flow: Most models rely on passive diffusion or non-physiological flow patterns (e.g., gravity-driven or pulsatile peristaltic pumps). They fail to replicate the continuous, high-magnitude intraluminal flow found in human organs, which is critical for nutrient delivery, signaling molecule transport, and mechanotransduction.
• Species Specificity and Control: Existing models often rely on animal grafting (hijacking mouse circulatory systems) to achieve flow, which reintroduces animal-specific variables and limits experimental control. Furthermore, monolayer endothelial cultures cannot recapitulate the 3D architecture or the complex hemodynamic forces (shear stress) required to study vascular remodeling and diseases like Hereditary Hemorrhagic Telangiectasia (HHT).

There is a critical need for a fully human, scalable, and tunable in vitro platform that can generate physiological intraluminal flow to dissect how hemodynamic forces regulate endothelial cell states and vessel architecture.

2. Methodology: The VIVOS Platform

The authors developed VIVOS (Vascularized In Vitro Organ Systems), an electromechanical platform designed to couple perfused human vascular beds with tunable pumps.

• Hardware Design:

  • Microfluidic Chips: Fabricated from PDMS, these chips contain a "vascular bed compartment" connected to high- and low-pressure channels via porous membranes with 20-µm slits. The slits prevent hydrogel seepage and minimize flow recirculation vortices.
  • Impeller Pumps: Instead of peristaltic or syringe pumps, VIVOS uses magnetically coupled rotating impeller pumps. These generate continuous, non-pulsatile flow with precise pressure differentials (up to ~610 Pa) and flow rates that mimic physiological circulation. The pumps are housed in "motor boxes" inside the incubator, allowing for sterile, connector-free removal of culture plates.
  • Scalability: The system supports parallelization, with one motor box controlling three 6-well plates simultaneously.

• Biological Components:

  • Vascular Beds: Generated by suspending human primary cells (Brain Microvascular Endothelial Cells - BMVECs, Normal Human Lung Fibroblasts - NHLFs, and Human Brain Vascular Pericytes - HBVPs) in a fibrin-based hydrogel. These self-assemble into robust vascular networks within 4–8 days.
  • Organoid Integration: The vascular beds are large enough (millimeter-scale) to accommodate diverse human organoids (lung, cerebral, vascular, breast spheroids) and retinal explants, allowing for the study of organ-specific vascular interfaces.

• Experimental Controls:

  • Flow Tuning: The system allows for "Low Flow" (300 RPM, 65 Pa) and "High Flow" (900 RPM, ~610 Pa) conditions, generating physiological shear stresses (2–3 dyn/cm²).
  • Disease Modeling: The platform was used to model HHT by knocking down ENG (Endoglin) or ALK1 in endothelial cells and observing the formation of arteriovenous malformations (AVMs) under defined flow.

3. Key Contributions

1. First Fully Human Flow-Quantified System: VIVOS is the first platform to provide continuous, tunable, physiological intraluminal flow in a fully human 3D tissue context without animal grafting.
2. Mechanistic Link Between Flow and Gene Regulation: The study establishes a direct molecular link between laminar shear stress and endothelial cell state transitions via a YAP/TAZ-Apelin signaling switch.
3. Disease Modeling: It successfully models HHT-associated AVMs in vitro, demonstrating that BMP9 signaling constrains vessel caliber and antagonizes VEGF-driven angiogenesis in a flow-dependent manner.
4. Transport Dynamics: The platform proves that intraluminal flow is the dominant mechanism for compound transport in millimeter-scale tissues, overcoming the diffusion limits of standard organoid cultures.

4. Key Results

A. Physiological Flow Drives Transport and Maturation

• Transport Mechanism: Quantitative analysis using fluorescent dextran showed that in VIVOS, compound transport is driven primarily by intraluminal vascular flow, not diffusion. Without flow, tissues larger than a few hundred microns fail to form vessels and undergo apoptosis (high cleaved caspase-3).
• Organoid Integration: VIVOS successfully vascularized diverse tissues (brain, lung, breast, retina). Cerebral organoids formed neurovascular units, and lung organoids integrated vessels adjacent to epithelium, mimicking in vivo interfaces.

B. The YAP/TAZ-Apelin Switch

Single-cell RNA sequencing (scRNA-seq) of vascular beds under low vs. high flow revealed a profound remodeling program:

• Shear Stress Sensing: High flow (physiological shear stress) suppresses YAP/TAZ activity (confirmed by reduced nuclear localization and target gene expression).
• The Switch:
  • Low Flow (High YAP/TAZ): Endothelial cells upregulate Apelin (APLN) and downregulate its receptor (APLNR). This state correlates with Tip Cell phenotypes (sprouting, angiogenic).
  • High Flow (Low YAP/TAZ): Endothelial cells downregulate APLN and upregulate APLNR. This state correlates with Stalk Cell phenotypes (stabilization, lumen formation).
• Functional Validation:
  • Inhibiting YAP/TAZ (using IAG933) mimicked high flow, reducing vessel sprouting (branchpoint density).
  • Inhibiting Apelin signaling (using ML221) increased sprouting, mimicking low flow.
  • Conclusion: Physiological laminar flow suppresses YAP/TAZ, shifting endothelial cells from a sprouting (Tip/APLN-high) state to a stabilizing (Stalk/APLNR-high) state. This acts as a negative feedback loop to prune excess vessels.

C. Modeling Hereditary Hemorrhagic Telangiectasia (HHT)

• AVM Phenotype: Knocking down ENG or ALK1 resulted in enlarged vessels with significantly higher flow velocities, recapitulating the AVM phenotype seen in HHT patients.
• BMP9 vs. VEGF:
  • BMP9 treatment constrained vessel diameter and reduced perfusion.
  • Mechanism: BMP9 signaling was found to antagonize VEGF-induced transcription. While VEGF upregulates angiogenic markers (e.g., ANGPT2, CXCR4), BMP9 suppresses these genes.
  • Dysregulation: In HHT models (loss of ENG/ALK1), this BMP9-mediated suppression of VEGF signaling is lost, leading to uncontrolled angiogenesis and enlarged vessels.
  • Therapeutic Insight: Treatment with the VEGF receptor inhibitor Axitinib rescued the enlarged vessel phenotype in HHT models, validating the antagonistic relationship between BMP9 and VEGF pathways.

5. Significance

• Preclinical Translation: VIVOS provides a "fully human" alternative to animal models for studying vascular development and disease, addressing the high failure rate of translating animal vascular findings to humans.
• Mechanistic Insight: The discovery of the YAP/TAZ-Apelin switch offers a new paradigm for understanding how hemodynamic forces dictate vascular architecture, moving beyond simple nutrient delivery to active mechanotransduction.
• Therapeutic Development: The platform enables the testing of flow-regulated pathologies (like HHT) and the screening of therapeutics (e.g., VEGF inhibitors) under physiologically relevant hemodynamic conditions.
• Future Applications: The modular design allows for multi-organ interconnection and the introduction of circulating immune cells, paving the way for studying metastasis, immune trafficking, and complex multi-organ diseases in a controlled in vitro environment.