Deep parallel characterization of AAV tropism and AAV-mediated transcriptional changes via single-cell RNA sequencing

Abstract

Engineered variants of recombinant adeno-associated viruses (rAAVs) are being developed rapidly to meet the need for gene-therapy delivery vehicles with particular cell-type and tissue tropisms. While high-throughput AAV engineering and selection methods have generated numerous variants, subsequent tropism and response characterization have remained low throughput and lack resolution across the many relevant cell and tissue types. To fully leverage the output of these large screening paradigms across multiple targets, we have developed an experimental and computational single-cell RNA sequencing (scRNA-seq) pipeline for in vivo characterization of barcoded rAAV pools at unprecedented resolution. Using our platform, we have corroborated previously reported viral tropisms and discovered unidentified AAV capsid targeting biases. As expected, we observed that the tropism profile of AAV.CAP-B10 in mice was shifted toward neurons and away from astrocytes when compared with AAV-PHP.eB. Our transcriptomic analysis revealed that this neuronal bias is mainly due to increased targeting efficiency for glutamatergic neurons, which we confirmed by RNA fluorescence in situ hybridization. We further uncovered cell subtype tropisms of AAV variants in vascular and glial cells, such as low transduction of pericytes and Myoc+ astrocytes. Additionally, we have observed cell-type-specific responses to systemic AAV-PHP.eB administration, such as upregulation of genes involved in p53 signaling in endothelial cells three days post-injection, which return to control levels by day twenty-five. Such ability to parallelize the characterization of AAV tropism and simultaneously measure the transcriptional response of transduction will facilitate the advancement of safe and precise gene delivery vehicles.

Adeno-associated virus (AAV) reduces cortical dendritic complexity in a TLR9-dependent manner

“…Four days after injection of AAV8-hSyn-mCherry, mRNAs encoding H2-K, H2-D, and C3 are upregulated in somatosensory cortex. These findings from mouse cortex align well with previous studies showing upregulation of MHCI four days after striatal injection of AAV9 or AAV2 in rats 50,51 , and AAV-induced upregulation of H2-D in cortical microglia of mice three days post-injection 52 . In addition to the rapid response, we see persistent upregulation of H2-K and H2-D proteins three weeks after injection of either AAV8-hSyn-mCherry or AAV1-pCAG-FLEX-EGFP in somatosensory cortex, and in AAV8-hSyn-mCherry-injected motor cortex and cerebellum.…”

Spatial Atlas of the Mouse Central Nervous System at Molecular Resolution

“…Next, we asked how AAV-PHP.eB tropism varies across molecular cell types. In addition to recapitulating the known tropism of PHP.eB towards neurons and astrocytes 5,15 (Extended Data Fig. 9c), we uncovered new viral tropism features across fine molecular cell types at unprecedented resolution.…”

“…In the hippocampal formation, molecular tissue regions correspond to CA1-3 and dentate gyrus (DG) in the hippocampal region as well as subiculum (SUB) and medial entorhinal area (ENTm) in the retrohippocampal region (Fig. 2d, slices [1][2][3][11][12][13][14][15]. Likewise, in the striatum, molecular tissue regions recapitulate known anatomical divisions, including the dorsal (region STR_A, D1 medium spiny neurons MSN_3 and MSN_4 enriched) and ventral (region STR_B, D1 MSN_2 and D2 MSN_7 enriched) areas, and islands of Calleja (regions isl_A and isl_B, D1 MSN_12 and MSN_9 enriched, respectively) (Fig.…”

“…Interestingly, we identified a core-shell-like structure in the retrosplenial cortex (RSP), presubiculum (PRE), and postsubiculum (POST), where layer 2 (region L2/3_E, marked by Tshz2 + Dkk3 + TEGLU_10) is surrounded by layers 1 and 3 (region L2/3_D, marked by Tshz2 + Cbln1 + TEGLU_35 or Tshz2 + Rxfp1 + TEGLU_30) (Fig. 2d, slices [12][13][14][15]. On the other hand, one molecular tissue region could contain multiple anatomically defined tissue regions, which suggests shared cell-type compositions.…”

“…One of its variants, PHP.eB, can efficiently cross the blood-brain barrier (BBB) after systemic administration, enabling brain-wide transgene expression and gene therapy development 5, 6, 14 . However, the cell- type and tissue-region tropisms of AAV-PHP.eB have not been resolved nor registered in the context of the spatial brain atlas, due to the lack of robust and scalable viral characterization strategies among all the cell types and tissue regions 15 .…”