A Compact Base Editor Rescues AATD-associated Liver and Lung Disease in Mouse Models

Abstract

Alpha-1 antitrypsin deficiency (AATD) is commonly caused by a G-to-A mutation in the SERPINA1 gene (the PiZ mutation). The mutant PiZ AAT protein is sequestered in hepatocytes, causing lung emphysema due to insufficient AAT protein to inhibit neutrophil elastase in the lung. Here we show that a compact adenine base editor (ABE) with an evolved Cas9 nickase derived from Neisseria meningitidis (eNme2.C) can be packaged in a single AAV and correct the PiZ mutation in mouse models of AATD. An all-in-one eNme2.C-TadA8e/guide 2 plasmid achieved approximately 20% on-target editing in PiZ reporter cells. TadA9e, which has a narrower editing window than TadA8e, reduced bystander editing without significantly affecting the on-target edit. In PiZ transgenic mice, eNme2.C-TadA9e AAV showed approximately 23% editing efficiency after 8 weeks and reduced liver disease burden in treated mice. In a new AAT-null;PiZ transgenic mouse model, ABE restored serum levels of AAT to beyond the 570 μg/mL therapeutic level. Moreover, ABE treatment was able to significantly correct lung functions in AAT-null;PiZ animals with emphysema. This study demonstrates the feasibility of an eNme2.C-based ABE in a single AAV to treat both AATD-associated liver and lung disease.

Keywords: AAV; Base editing; gene editing; mouse liver; rare diseases.

Figure 1.. Optimization of an all-in-one editor for AAT correction.

(a) PiZ mutation site with the pathogenic A mutation noted in red. Amino acids are in bold. (b) Two guides were tested and delivered to PiZ reporter cells. PAM and editing window for each guide are shown. (c) A to G editing efficiency of the two guides in PiZ reporter cells. Numbers represent mean, n=3. Target A is in red, with potential bystander edits in blue. (d) Editing efficiency of all-in-one construct that can deliver eNme2.C-TadA8e. Mean ± SD shown, n=3. (e) Mutant-specific sgRNA edits PiZ reporter but not the wildtype PiM with a single G mismatch. Red arrow depicts target base while blue depicts bystanders. Representative image shown.

Figure 2.. Reduction of bystander edits with TadA9e.

(a) Editing efficiency of all-in-one construct that can deliver eNme2.C-TadA9e in reporter cells. Mean ± SD shown, n=3. (b) Ratio of bystander to on-target edit in PiZ reporter cells. Mean ± SD shown. *p ≤ 0.05, **p ≤ 0.01,***p ≤ 0.001 by unpaired t-test. (c) Allele population of cells edited with eNme2.C-TadA8e or eNme2.C-TadA9e. Line depicts median, n=3. (d) Deep sequencing of OT sites. Mean ± SD shown. See Supplemental file 1 for additional information, n=3.

Figure 3.. In vivo editing of PiZ animals.

(a) Schematic of tail-vein injection in PiZ animals with all-in-one eNme2.C-TadA9e. (b) eNme2.C-TadA9e in vivo editing at 4 and 8 weeks in PiZ animals. Numbers represent mean (n=3 for control, n=5 for 4 weeks, and n=7 for 8 weeks). (c) Allele percentage in PiZ mice treated with eNme2.C-TadA9e. Line represents median, n=5 for 4 weeks, n=7 for 8 weeks. (d) Representative images of mouse liver stained with PAS-D. Mice were either untreated or treated with 9e11vg AAV for 8 weeks. Scalebar: 500μm for top images, 100μm for bottom images. Squares represent area of zoom from the top images, while circle represents the PASD+ and PAS− regions. Right: quantification. Mean ± SD shown. Significance was calculated by unpaired t-test, n=3.

Figure 4.. Lung disease reversal in AAT-null;PiZ animal.

(a) Schematic of AAT-null;PiZ mouse generation (b) Schematic of ABE treatment to study LPS-induced emphysema in AAT-null;PiZ animals. (c) Editing in AAT-null;PIZ animals when treated with 9e11vg AAV for 15 weeks (n=4 for control, n=6 for ABE) (d) AAT levels in control and treated AAT-null;PiZ animals. Mean ± SD shown, n=3 for control, n=5 for ABE. Both groups were treated with LPS. (e-i) Inspiratory capacity (e), compliance (f), elastance (g), resistance (h) and (i) pressure-volume (PV) loop of PBS, LPS, or ABE+LPS treated animals (n=7, n=10, and n=9, respectively). Line depicts median. Significance was detected by one-way ANOVA with uncorrected Fisher’s least significance difference, except for PV-loop (two-way ANOVA with Dunnett. Mean ± SEM shown), *p ≤ 0.05,****p ≤ 0.0001.