Expansion in the number of AAV serotypes, both through the identification of novel natural serotypes and new, engineered serotypes [16, 23, 25–35], has resulted in improved gene transfer to specific cell types in vivo. There have been several publications reviewing the preferred in vivo tropism of these new serotypes [2, 36]. However ex vivo/in vitro data is lacking. Alternative uses of AAV include using this viral vector ex vivo as a method of gene transfer into specific cell types. These transduced cells could then be studied directly or used for cell-based gene therapy, whereby AAV transduction would occur and the modified cells would then be transplanted.
Another potential use of AAV vector transduction ex vivo would be to use AAV vectors to stimulate gene targeting by homologous recombination, to precisely modify the genome of the cells to be transplanted. This modification could be done through gene targeting directly by AAV [37–40] or in combination with the induction of a site-specific double-strand break. These site-specific double-strand breaks could be induced by a homing endonuclease [21, 41, 42], by zinc finger nucleases , or by some other nuclease, like TAL effector nucleases (TALENs) [44–47]. An important aspect to using AAV in this manner is to determine the best serotype to transduce specific cell types ex vivo, where there is no basement membrane or extracellular matrix. In fact, we have already reported the ability of a single AAV6 vector to deliver both zinc-finger nucleases as well as a donor repair substrate, to stimulate gene targeting .
In regards to stem cells, it is intriguing that many progenitor cells did not transduce well (i.e. see lots of dark blue in the heat maps). Perhaps the cells have evolved the ability to avoid transduction as a way to protect themselves from changes to the cell, in particular the DNA. However, there certainly were stem cells that were transduced well by various AAV serotypes. This again points to the utility of this study, as certain serotypes were good for transduction in some stem cells and bad in others and although AAV1 and AAV6 were good or the best at transduction in many stem cells, there were some stem cells that transduced poorly with AAV1 and AAV6.
In this work we provide a broad survey that examines the ability of ten different AAV serotypes to infect thirty-four different cell types ex vivo/in vitro. In general, we demonstrate that AAV1 and AAV6 have the greatest ability to transduce a wide range of cell types. We found, however, that for particular cell types there are specific serotypes, which provide optimal transduction. (For example, AAV4 is the optimal serotype for transducing murine adipose progenitor cells.) We also found that there are certain primary cell types, such as human hematopoietic progenitor cells, that were not efficiently transduced by any of the ten different serotypes. It is possible that the lack of measured transduction in these cell types is because the CMV promoter is relatively weak in these cells. It is not likely, however, that in these cases the cells were overloaded with uptake and processing of virions because when a 10-fold lower MOI was used, lower transduction efficiency was seen in every case (data not shown). Although, different growth conditions were used for many of the different cell types, each serotype was used for each cell type in the same conditions, thus providing an internal control for a comparative analysis. However, because a small amount of our data does not perfectly match with previous findings, we suggest that the results presented here should lead investigators to choose a few of the best serotypes for their specific need. Our results demonstrate that there is no simple mechanism to predict which serotype will transduce a particular cell type but does suggest that if one were limited to screening a small number of serotypes that focusing on AAV6, AAV2, and AAV3 would be reasonable as those three serotypes give a broad range of effectiveness across most cell types. In the future, it may be important to further study the transduction of AAV after different purification strategies are used, as it has been shown to affect transduction . Methods to facilitate AAV transduction, such as by the use of proteasome inhibitors  or strategies that allow for selection of novel capsids, may help overcome the barrier to transduction that these cells exhibit. However, it is likely, there are factors, unproven as of yet, that serve as major barriers to transduction by AAV. For example, it is possible that the apparent low transduction could be a consequence of AAV vectors inducing apoptosis . In this case, a caspase inhibitor such as Z-VAD-FMK could be used to achieve transduction without cell death. Understanding these unproven barriers to transduction would further improve the utility of AAV as a gene transfer vector for ex vivo manipulation of primary cells as well as in vivo gene therapy.
In summary, we have performed a survey of the ability of different AAV serotypes to transduce a wide variety of different primary and immortalized cell types. This survey should be a useful and practical resource for investigators as they consider using AAV as a gene transfer vector in their studies.