Targeting the somatosensory system with AAV9 and AAV2retro viral vectors

Abstract

Adeno-associated viral (AAV) vectors allow for site-specific and time-dependent genetic manipulation of neurons. However, for successful implementation of AAV vectors, major consideration must be given to the selection of viral serotype and route of delivery for efficient gene transfer into the cell type being investigated. Here we compare the transduction pattern of neurons in the somatosensory system following injection of AAV9 or AAV2retro in the parabrachial complex of the midbrain, the spinal cord dorsal horn, the intrathecal space, and the colon. Transduction was evaluated based on Cre-dependent expression of tdTomato in transgenic reporter mice, following delivery of AAV9 or AAV2retro carrying identical constructs that drive the expression of Cre/GFP. The pattern of distribution of tdTomato expression indicated notable differences in the access of the two AAV serotypes to primary afferent neurons via peripheral delivery in the colon and to spinal projections neurons via intracranial delivery within the parabrachial complex. Additionally, our results highlight the superior sensitivity of detection of neuronal transduction based on reporter expression relative to expression of viral products.

Fig 1. Targeting of spinoparabrachial neurons by AAV9, AAV2retro, and DiI.

A-C, Representative images of the parabrachial complex (A1, B1, C1; scale bars = 500 μm) and dorsal spinal cord (A2, B2, C2; scale bars = 50 μm) following parabrachial complex injections of AAV9 (A), AAV2retro (AAV2r, B), or Fast-DiI (C). The images in B2 and C2 represent a maximum intensity projection of 3D stacks spanning 12 and 16 μm, respectively, and collected with a 4-μm z-step. Arrowheads in A2, B2 and C2 highlight labeled spinal projection neurons in the superficial dorsal horn. Additionally, in B2, the short arrow highlights a plexus of labeled processes in lamina I, and the long arrow indicates labeling of the dorsal corticospinal tract. D, Stacked Bars represent the relative distribution of labeled projection neurons between the dorsal horn (DH) and the lateral spinal nucleus (LSN) in the ipsilateral (left) and contralateral (right) spinal cord following injection with AAV9 (top), AAV2retro (middle), or Fast-DiI (bottom) (** p<0.01, **** p<0.0001, compared to Fast-DiI; $ $ p<0.01, compared to AAV2r). The total number of labeled neurons analyzed and the number of animals (n) is noted under the stacked bars. Pie charts in D show the percentage of labeled neurons present in lamina I-V of DH. Proportions of labeled neurons in spinal laminae were compared using Two-way ANOVA and Bonferroni posttest ($, p<0.05 compared to AAV2r). E, The average labeling of spinal tracts at the level of L3 is shaded for AAV9 (E1) and AAV2r (E2). F, Retrograde labeling of cortical neurons in primary motor cortex (M1) following parabrachial injection of AAV2retro; scale bar = 150μm.

Fig 2. Intraspinal delivery of AAV9 and AAV2retro viral vectors leads to effective transduction of both spinal and primary afferent neurons.

A-D, tdTomato (red, subscript 1) and GFP (green, subscript 2) immunofluorescence in Ai14 mice injected unilaterally in the L3/L4 region of spinal cord with AAV9 (A and B), or AAV2retro (C and D); scale bars = 150 μm. The discrepancy in the tdTomato and GFP labeling is largely due to the differential subcellular localization of the two reporter proteins, and more specifically to the overwhelming tdTomato labeling in the neuropil of densely transduced dorsal horn regions and in the central processes of transduced DRG neurons. The high fluorescence intensity of these regions required imaging parameters that were not optimal for regions with lower density of transduced neurons. In contrast, the discrete nuclear localization of Cre-GFP allowed clear visualization of individual nuclei in the spinal cord. E and F, tdTomato (red, subscript 1) and GFP (green, subscript 2) immunofluorescence in neurons of the ipsilateral L4 DRG following transduction with AAV9 (E) or AAV2retro (F); not all tdTomato-labeled cells were immunoreactive for nuclear GFP, scale bars = 150μm. G and H, NeuroTrace staining (blue, subscript 1), tdTomato (red, subscript 2) and CGRP (green, subscript 3) immunofluorescence in the ipsilateral L4 DRG following transduction with AAV9 (G) or AAV2retro (H); arrows point to examples of tdTomato and CGRP colocalization, scale bars = 100 μm.

Fig 3. Transduction of spinal cord and DRG neurons following intrathecal delivery of AAV9 and AAV2retro.

A and B, tdTomato (red, subscript 1) and GFP (green, subscript 2) immunofluorescence in the lumbar spinal cord of mice injected with AAV9 (A) or AAV2retro (B); scale bars = 200μm. C and D, tdTomato (red, subscript 1) immunofluorescence and NeuroTrace staining (blue, subscript 2) showing transduced neurons in L4 DRG following injection of AAV9 (C) or AAV2retro (D); scale bars = 100μm. E, NeuroTrace (C2, D2) was used to quantify the area of individual L4 DRG neurons. The areas of total (E1), AAV9 transduced tdTomato+ (E2), and AAV2retro transduced tdTomato+ (E3) DRG neurons are shown as frequency histograms to illustrate the number of DRG neurons relative to their area (μm²). F, The cumulative frequency distribution plot of DRG neuron area shows that while the size distribution of AAV9 tdT+ DRG neurons (red line) closely matches that of total NeuroTrace labeled DRG neurons (black line; Kolmogorov-Smirnov test, p>0.05), the size distribution of the AAV2retro tdT+ DRG neuron population (blue line) is skewed toward larger cells compared to both the total DRG neuron population, and the AAV9 tdT+ DRG neuron population, with 66.35% of the total DRG neuron population and 59.84% of the AAV9 tdT+ DRG neuron population comprised of neurons with an area ≤ 500 μm², while only 33.06% of the AAV2r+ DRG neuron population has an area ≤ 500 μm² (Kolmogorov-Smirnov test, **** = p<0.0001). G, Each dot represents the mean area of DRG neurons from a single animal (n = animal #). The mean area of AAV2retro tdT+ DRG neurons is significantly larger than the mean area of total, and AAV9 tdT+, DRG neurons (Welch’s one-way ANOVA with Dunnet’s T3 multiple comparisons test ** = p<0.01; **** = p<0.0001).

Fig 4. AAV9 and AAV2retro-mediated transduction following intracolonic injections.

A and B, tdTomato labeling in L6 DRG and sacral spinal cord was more abundant in AAV9-injected compared to AAV2retro injected animals from cohort 1. The images in A1 and B1 represent a maximum intensity projection of four stitched 3D stacks spanning the entire ganglion (collected with a 3 μm z-step). Scale bars: A1 and B1, 150 μm; A2 and B2, 100 μm. C, Quantitative analysis of tdTomato expressing L6 DRG neurons from cohort 1 (circles) and cohort 2 (squares). D-E, Sacral spinal cords from cohort 2 showed labelling of afferent fibers and a few neurons in the spinal cord, consistent with the previously observed pattern (D, E). Scale bar = 100μm. F-G, The presence of tdTomato labelled fibers and cells in cervical spinal cord as well as cells of the liver provides evidence for systemic distribution of AAV9 (F, G). Scale bars: F, 100μm; G, 150μm.