Systemically injected AAV virions recirculate through the bloodstream until a capsid surface loop binds a cell-surface receptor and the particle is captured (transduces or is scavenged), or it is cleared by the kidneys / immune opsonization. On each pass through an organ bed the probability of capture is a saturable receptor-binding term set by the serotype's affinity for that tissue's receptors:
P(capture at organ i) = κ_i(serotype) — per-pass capture probability
Route choice at the circulation hub ∝ Q_i — organ blood-flow fraction (cardiac output share)
P(still circulating after n passes) = Π (1 − κ_i − c) — c = per-pass clearance rate
- Serotype — sets κ_i for each organ. AAV8 binds hepatocyte receptors so strongly that most systemic dose is sequestered by the liver before it ever reaches a distant target. AAV9/AAVrh10 partially cross the blood-brain barrier (drawn as a translucent shell around the CNS node); an engineered, liver-detargeted capsid trades some potency everywhere for a much lower liver κ.
- Target tissue — which organ is being treated; the "on-target capture" readout tracks only that node.
- Injection route — systemic (IV) sends every particle through the full circulation, giving the liver first crack at high-affinity serotypes; local (intramuscular / intrathecal / subretinal-style direct) delivery puts the first pass straight at the target organ, which is exactly why real AAV gene therapies (e.g. intrathecal AAV9 for SMA, subretinal AAV2 for inherited retinal disease) use local routes to dodge hepatic sequestration.
- A particle that is not captured and not cleared loops back through the circulation hub and gets re-routed by organ blood flow again — this is why some particles visibly make several passes before settling.
Real-world relevance: hepatic sequestration of systemic AAV is one of the central dose-limiting problems in gene therapy — it wastes vector, drives the liver toxicity seen at high IV doses, and is a major reason capsid engineering and local delivery routes exist.