AAV Frequently asked questions
Engineered capsid use and biodistribution
We and others have found that AAV-PHP.B and PHP.eB work in many mouse (Fig. 1E) strains and rats (Fig. 5). Three independent reports have confirmed that the principal receptor for the AAV-PHP.B capsid family (PHP.B/PHP.eB/PHP.V1/PHP.N) is the protein Ly-6a, which is polymorphic across strains (See Supplementarty Table 3). Please confirm that your strain of interest contains the appropriate form of Ly-6a before proceeding with the above capsids.
AAV-PHP.B appears to provide increased CNS transduction, relative to AAV9, in young rats, and in collaboration with Nina Huber and Sergiu Pasca we showed that AAV-PHP.B transduces human neurons and glia in cortical spheroids (see Supplementary Fig. 5). Consistent with reports identifying the murine-restricted protein Ly-6a as the principal receptor for the AAV-PHP.B capsid family (PHP.B/PHP.eB/PHP.V1/PHP.N), we have heard from several groups that AAV-PHP.B does not transduce the CNS of songbirds with increased efficiency, following IV injection. This has also been reported for PHP.B in marmoset.
AAV-PHP.eB transduces the CNS as well as many other organs. Please see Deverman et al 2016 for biodistribution data and images of several organs following PARS-based tissue clearing.
Broad off-target transduction (generally consistent with AAV9) is also seen with PHP.S and PHP.V1.
PHP.N and CAP-B10 both show markedly reduced transduction outside the CNS. Please see capsid-relevant publications for details.
With shipping especially international shipping, it is highly advised to re-titer before injection to animals. Also it would be the best to not refreeze, as each freeze/refreeze degrades the titer. Virus is stable for many weeks/months at 4 ̊C, however, we recommend that users re-titer if it has been more than 2 months. More information can be found in the paper here.
You will likely have to test this experimentally.
We typically administer between 1 × 1011 and 5 × 1011 vg of AAV-PHP.eB or between 3 × 1011 and 1 × 1012 vg of AAV-PHP.S to adult mice (6–8 weeks old). However, dosage will vary depending on the target cell population, desired fraction of transduced cells, and expression level per cell. AAVs independently and stochastically transduce cells, typically resulting in multiple genome copies per cell. Therefore, higher doses generally result in strong expression (i.e., high copy number) in a large fraction of cells, whereas lower doses result in weaker expression (i.e., low copy number) in a smaller fraction of cells.
To achieve high expression in a sparse subset of cells, users can employ a two-component system in which transgene expression is dependent on co-transduction of an inducer (e.g., a vector expressing Cre, Flp, or the tetracycline-controlled transactivator (tTA)); inducers are injected at a lower dose (typically 1 × 109 to 1 × 1011 vg) to limit the fraction of cells with transgene expression. Note that transgenes and gene regulatory elements (e.g., enhancers, promoters, and miRNA target sequences) can influence gene expression levels.
Therefore, users should assess transgene expression from a series of doses and at several time points after intravenous delivery to determine the optimal experimental conditions.
Not to our knowledge. We use retro-orbital sinus injections on the advice of our veternarian staff and because it's easier and more reliable in our hands. Based on data from other groups, tail vein injection appears to provide similar vector distribution.
We access the vasculature in mice via the retro-orbital sinus. Please see this helpful paper for a protocol and a discussion of the benefits of this route versus tail vein injections.
The expression time will vary depending on the strength of the promoter and the level of transgene expression required. With a strong ubiquitous promoter like CAG, two weeks is often sufficient. For weaker promoters or high expression transgenes, it can be up to 4 weeks. Therefore, users should assess transgene expression from a series of doses and at several time points after intravenous delivery to determine the optimal experimental conditions.
We strongly recommend using AAV-PHP.eB. Together with the GfABC1D promoter developed by Michael Brenner (U. of Alabama), AAV-PHP.eB will provide efficient and selective gene expression in astrocytes (see efficiency data in Chan et al 2017).
AAV-PHP.A from Deverman et al 2016, while more selective for astrocytes, requires higher virus doses and is extremely difficult to work with because the virus produces poorly (AAV-PHP.A yields are typically ~30X lower than AAV9) and because it also aggregates quickly in PBS. We have not figured out how to store this virus without significant (visible) aggregation and do not recommend its use.
AAV-PHP.eB and AAV9 appear to spread to a similar degree based on injections into the mouse striatum. In our unpublished data, there was no obvious difference in tropism or spread when injected directly into the brain.
AAV-PHP capsids do not differ from other natural AAV capsids in this regard. We have achieved high titer preparations of genomes around 4.7kb. As expected, genomes larger than 5kb have not packaged well.
The PHP.B serotype is slightly more prone to loss (possibly due to aggregation) during the prep as compared to AAV9. This should not be a problem if the purification is done following the protocol which we provide. It may work with other protocols, but we cannot guarantee this.
Once purified, AAV-PHP.B can be concentrated to very high titers (>1e14 vg/ml) and the yields are similar to AAV9. Note, we lyse the cells in a high salt buffer. While this step may not be essential, it may reduce the risk of aggregation during purification.
Yes, use a standard AdV helper.
Yes, use AAV2 ITRs. We do not find it necessary to use scAAV genomes to achieve rapid, efficient transduction of the CNS via the vasculature. We have also successfully produced scAAV genomes. Be aware however that we and others have observed shifts in vector tropism for a given engineered capsid when packaging scAAV genomes.
The iCAP-PHP plasmids have a tTA-TRE based inducible amplifiction loop built in to increase virus production. In our hands, these iCAP plasmids increase virus production by nearly 2-fold as compared to standard rep-cap plasmids when used at the transfection ratios provided in our protocol. This is a tet-off system, so no dox or tet is needed to turn it on and it can be used like any other rep-cap plasmid. *This system should only present a problem if the rAAV genome to be packaged has a tet responsive element that directs expression of a protein that affected the health of the production cells. For some AAV-PHP variants, we can provide the capsid in a standard rep-cap plasmid format.
Do not inject more than 10% of the mouse blood volume, which corresponds to 150 µl for a 25-g mouse.
Depending on the user, it is easiest to inject 40–80 µl/mouse. If <40 µl/mouse is required, use DPBS or saline to dilute the virus such that a larger volume is injected. If more than 80 µl/mouse is required, it may be more convenient to re-concentrate the virus or perform two separate injections; follow institutional guidelines for multiple eye injections. Virus will be lost in the event of an unsuccessful injection; therefore, prepare more master mix than is required.
This varies from prep to prep, and from genome to genome. An optimal yield post purification is 2-4E12 vg/150 mm dish of producer cells.
We've never tested lipofectamine. Its use would add significant cost to this procedure given the scale of transfection. Also, there may be something particularly advantagous about PEI transfections because the transfection efficiency of 293 cells is lower than with Ca2+ phosphate, but the virus yields are nearly 10x higher, at least in our hands.
We use a genome: capsid: helper ratio of 1:4:2 based on ug of DNA. We use 40 ug of DNA total for each 150 mm dish. Supplemental Table 2 from our Nature Protocols paper is a transfection calculator to simplify the process.