Caveolae-mediated transcytosis vs. endocytosis: why active transport wins
Most tissue-targeting strategies in drug delivery rely, whether they admit it or not, on one of two mechanisms: passive accumulation (drugs leaking into tissue through gaps in poorly formed blood vessels — the EPR effect many nanoparticle and ADC designs exploit) or receptor-mediated endocytosis (a cell-surface receptor pulls cargo inside the cell, where much of it is degraded before ever reaching the far side of the vessel wall).
Caveolae-mediated transcytosis is neither. It's an active, energy-dependent pumping mechanism that moves cargo through the endothelial cell and out the other side — and the evidence for why that distinction matters isn't theoretical.
Consider what happens when a caveolae-targeting antibody is engineered against truncated Annexin A1, a protein enriched on the tumor endothelial surface. In vivo, the antibody reaches roughly 15% injected dose per gram in tumor tissue within one hour — more than 100 times the uptake of a non-targeting antibody — while normal organs see under 1% (Oh et al., Nature Medicine, 2014). That's not a modest improvement over passive accumulation; it's a different mechanism producing a different order of magnitude of result. Tumor signal exceeds peak blood concentration within the same hour — the antibody isn't accumulating opportunistically, it's being actively concentrated against the gradient.
The lung tells the same story from a different angle. An antibody targeting aminopeptidase P2, enriched in lung microvascular caveolae, achieves first-pass lung extraction exceeding 80%, with lung concentration reaching roughly four times peak blood levels (Chrastina et al., J Vascular Research, 2010). Critically, when caveolin-1 — the structural protein caveolae are built from — is experimentally knocked down, this transport is abolished. Not reduced. Abolished. That single experiment is what separates a real active-transport mechanism from one that merely correlates with a protein's presence: remove the pump, and the cargo stops moving, no matter how well it binds its target.
Why "active" is the axis most platforms miss
A protein can be highly disease-relevant and easy to bind, and still be a poor delivery target if the only way to reach it is passive diffusion through a leaky vessel — a mechanism that varies wildly between patients, tumor types, and even regions of the same tumor. An active transport route, once mapped and validated, behaves far more predictably.
This is precisely why AVATAR's scoring separates "druggable" from "accessible via active transport." A target can score high on the first axis and still fail the second — and in our experience, that gap is exactly where most preclinical delivery failures originate.
Read the original research: Oh et al., Nature Medicine (2014) · Chrastina et al., Journal of Vascular Research (2010)