Person:

Liu, David

Base editors are widely used engineered proteins that introduce targeted point mutations into DNA without creating double-stranded DNA breaks. The application of base editors to study and treat genetic diseases is dependent on their in vivo delivery into relevant cell types. Delivery by adeno-associated virus (AAV), a clinically validated delivery method, poses special challenges because the size of base editors exceeds the AAV packaging limit. Here we describe the development and application of in vivo delivery methods for both cytosine base editors (CBEs) and adenine base editors (ABEs). Dual AAVs each provide one half of the editor and trans-splicing inteins reconstitute full base editor activity, circumventing the AAV packaging limit. We optimized each AAV component to greatly improve editing efficiency. The resulting AAVs enable in vivo base editing for the first time in mouse brain, retina, and heart, as well as the most efficient base editing to date in muscle and liver, with therapeutically relevant efficiencies at viral dosages known to be tolerated in humans. A single intravenous injection of split CBE in PHP.eB AAV resulted in editing of up to 59% of unsorted mouse cortical tissue. Intravenous injection of split CBE or ABE in AAV9 mediated up to 38%, 20%, and 9% base editing in unsorted mouse liver, heart, and skeletal muscle, respectively. Subretinal and intracerebroventricular injections of split CBE and ABE in Anc80, PHP.B/PHP.eB, and AAV9 viruses mediated mouse retina and brain editing in up to 38% and 50% of unsorted cells, respectively. We applied this system to directly correct in mouse brain tissue a mutation that causes the neurodegenerative ataxia Niemann-Pick disease type C (NPC), slowing neurodegeneration and increasing lifespan consistent with expectations based on mosaic NPC mice. These findings establish a broadly useful AAV platform for the efficient introduction of targeted point mutations into multiple tissues of therapeutic interest for which in vivo base editing has not been previously reported.

Genome-wide association studies have uncovered hundreds of autoimmune disease-associated loci; however, the causal genetic variant(s) within each locus are mostly unknown. Here, we perform high-throughput allele-specific reporter assays to prioritize disease-associated variants for five autoimmune diseases. By examining variants that both promote allele-specific reporter expression and are located in accessible chromatin, we identify 60 putatively causal variants that enrich for statistically fine-mapped variants by up to 57.8-fold. We introduced the risk allele of a prioritized variant (rs72928038) into a human T cell line and deleted the orthologous sequence in mice, both resulting in reduced BACH2 expression. Naive CD8 T cells from mice containing the deletion had reduced expression of genes that suppress activation and maintain stemness. Our results represent an example of an effective approach for prioritizing variants and studying their physiologically relevant effects.

The all-protein cytosine base editor DdCBE uses TALE proteins and a double-stranded DNA-specific cytidine deaminase (DddA) to mediate targeted C•G-to-T•A editing. To improve editing efficiency and overcome the strict TC sequence-context constraint of DddA, we used phage-assisted non-continuous and continuous evolution to evolve DddA variants with improved activity and expanded targeting scope. Compared to canonical DdCBEs, base editors with evolved DddA6 improved mitochondrial DNA (mtDNA) editing efficiencies at TC by 3.3-fold on average. DdCBEs containing evolved DddA11 offered a broadened HC (H = A, C or T) sequence compatibility for both mitochondrial and nuclear base editing, increasing average editing efficiencies at AC and CC targets from less than 10% for canonical DdCBE to 15–30% and up to 50% in cell populations sorted to express both halves of DdCBE. We used these evolved DdCBEs to efficiently install disease-associated mtDNA mutations in human cells at non-TC target sites. DddA6 and DddA11 substantially increase the effectiveness and applicability of all-protein base editing.

Hutchinson-Gilford progeria syndrome (HGPS) is typically caused by a dominant-negative C•G-to-T•A mutation (c.1824 C>T, G608G) in LMNA, the nuclear lamin A gene. This mutation causes RNA mis-splicing that produces progerin, a toxic protein that induces rapid aging and shortens lifespan to ~14 years1-4. Adenine base editors (ABEs) perform targeted A•T-to-G•C base pair conversion with minimal byproducts and without requiring double-strand DNA breaks or donor DNA templates5,6. Here, we describe the use of an ABE to directly correct the pathogenic HGPS mutation in cultured progeria patient-derived fibroblasts and in a mouse model of HGPS. Lentiviral delivery of ABE to patient-derived fibroblasts results in ~90% correction of the pathogenic allele, mitigation of RNA mis-splicing, reduced progerin levels, and correction of nuclear abnormalities. Unbiased off-target DNA and RNA analysis did not detect off-target editing activity in treated patient-derived fibroblasts. In transgenic mice homozygous for the human LMNA c.1824 C>T allele, a single retro-orbital injection of adeno-associated virus 9 (AAV9) encoding the ABE resulted in substantial, durable correction of the pathogenic mutation (~20-60% across various organs 6 months post-injection), restoration of normal RNA splicing, and reduction of progerin protein. In vivo base editing rescued vascular pathology, preserving vascular smooth muscle cell counts and preventing adventitial fibrosis. A single ABE AAV9 injection at P14 improved animal vitality and greatly extended median lifespan from 215 to 510 days. These findings support the potential of in vivo base editing to treat HGPS, and other genetic diseases, by directly correcting the root cause of disease.

Most genetic diseases arise from recessive point mutations that require correction, rather than disruption, of the pathogenic allele to benefit patients. Base editing has the potential to directly repair point mutations and provide permanent therapeutic restoration of gene function. We developed a base editing strategy to treat Baringo mice, which carry a recessive, loss-of-function point mutation (c.A545G, resulting in the substitution p.Y182C) in transmembrane channel-like 1 (Tmc1) that causes deafness. Tmc1 encodes a protein that forms mechanosensitive ion channels in sensory hair cells of the inner ear and is required for normal auditory function. We found that sensory hair cells of Baringo mice have a complete loss of auditory sensory transduction that causes profound deafness. To repair the Baringo mutation, we tested several optimized cytosine base editors (CBEmax variants) and guide RNAs in Baringo mouse embryonic fibroblasts. We packaged the most promising CBE, derived from an activation-induced cytidine deaminase (AID), into dual AAV vectors using a split-intein delivery system. The dual AID-CBEmax AAVs were injected into the inner ears of Baringo mice at postnatal day 1. Injected mice showed up to 51% reversion of the Tmc1 c.A545G point mutation to wild type sequence (c.A545A) in Tmc1 transcripts. Repair of Tmc1 in vivo restored inner hair-cell sensory transduction, hair-cell morphology, and partial low-frequency hearing four weeks post-injection. These findings provide a foundation for a potential one-time treatment for recessive hearing loss and support further development of base editing to correct pathogenic point mutations.

Directed evolution can generate proteins with tailor-made activities. However, full-length genotypes, their frequencies and fitnesses are difficult to measure for evolving gene-length biomolecules using most high-throughput DNA sequencing methods, as short read lengths can lose mutation linkages in haplotypes. Here we present Evoracle, a machine learning method that accurately reconstructs full-length genotypes (R2 = 0.94) and fitness using short-read data from directed evolution experiments, with substantial improvements over related methods. We validate Evoracle on phage-assisted continuous evolution (PACE) and phage-assisted non-continuous evolution (PANCE) of adenine base editors and OrthoRep evolution of drug-resistant enzymes. Evoracle retains strong performance (R2 = 0.86) on data with complete linkage loss between neighboring nucleotides and large measurement noise, such as pooled Sanger sequencing data (~US$10 per timepoint), and broadens the accessibility of training machine learning models on gene variant fitnesses. Evoracle can also identify high-fitness variants, including low-frequency ‘rising stars’, well before they are identifiable from consensus mutations