Table 2.

Competing hypotheses explaining AAV vector immune responses in humans

Hypothesis	Supporting observations	Observations not supporting the model
Capsid antigen presentation and memory T-cell activation leads to clearance of AAV-transduced cells.67	Human studies show expansion of capsid T cells following vector delivery4,15,67;	Most humans are exposed to wild-type AAV, but not all react against the AAV capsid.
	Healthy subjects carry capsid-reactive T cells67,85;
	Capsid antigen is processed and presented on MHC I.91,95
Expression of rep/cap from vector impurities triggers CTLs against transduced cells.82	AAV capsid packaging DNA impurities can be found in vector preparations.84,85	Studies in mice failed to detect expression of Rep and Cap in animals injected with high vector doses.86
Translation of ARFs within expression cassettes results in CTLs against AAV-transduced cells.82	Animal studies show that it is possible to express epitopes from ARFs.87	Screening of subjects for anti-ARF T-cell reactivity did not support the model.4
Preferential uptake of AAV2 by APCs via heparin-binding domain results in higher immunogenicity.	Animal studies showed that AAV2 binds more efficiently to APCs than AAV8.88	Subjects dosed with AAV8 vectors (not binding to heparin) developed T-cell responses against the capsid.4

APC, antigen-presenting cell; ARF, alternate open reading frame.