Rapid and definitive treatment of phenylketonuria in variant-humanized mice with corrective editing

Abstract

Phenylketonuria (PKU), an autosomal recessive disorder caused by pathogenic variants in the phenylalanine hydroxylase (PAH) gene, results in the accumulation of blood phenylalanine (Phe) to neurotoxic levels. Current dietary and medical treatments are chronic and reduce, rather than normalize, blood Phe levels. Among the most frequently occurring PAH variants in PKU patients is the P281L (c.842C>T) variant. Using a CRISPR prime-edited hepatocyte cell line and a humanized PKU mouse model, we demonstrate efficient in vitro and in vivo correction of the P281L variant with adenine base editing. With the delivery of ABE8.8 mRNA and either of two guide RNAs in vivo using lipid nanoparticles (LNPs) in humanized PKU mice, we observe complete and durable normalization of blood Phe levels within 48 h of treatment, resulting from corrective PAH editing in the liver. These studies nominate a drug candidate for further development as a definitive treatment for a subset of PKU patients.

Fig. 1. Adenine base editing for correction of PAH P281L variant in human hepatocytes in vitro.

a Schematic of genomic site of the PAH P281L variant, adapted from UCSC Genome Browser (GRCh38/hg38). The vertical yellow bar indicates the G altered to A (in red) by the variant. The A two positions upstream of the variant (black arrow) is a potential site of bystander editing. The horizontal yellow bar and green bar indicate protospacer (thick) and PAM (thin) sequences targeted by PAH1 gRNA and PAH2 gRNA, respectively. b Generation of PAH P281L homozygous HuH-7 cell line. Top, sequence from wild-type HuH-7 cells. Middle, sequence from a pool of HuH-7 cells following prime editing. Bottom, sequence from an isolated clonal line homozygous for the P281L variant (HuH-7 P281L cells). Red arrows indicate the P281L site. c A-to-G editing following transfection of HuH-7 P281L cells with plasmids encoding ABE/gRNA combinations (n = 2 biological replicates). d A-to-G editing observed in dose-response studies with HuH-7 P281L cells treated with ABE8.8/PAH1 LNPs (left) or ABE8.8/PAH2 LNPs (right). On-target editing includes all outcomes with corrective P281L editing, irrespective of bystander editing (n = 3 biological replicates). e qRT-PCR from HuH-7 samples using flanking primers in PAH exons 7 and 8 and flanking primers in PAH exons 1 and 2, displaying the quantified ratio of exon 7/exon 8 to exon 1/exon 2 (n = 3 biological replicates; mean ± standard deviation for each condition, normalized to untreated wild-type HuH-7 cells). f On-target or off-target editing at top ONE-seq-nominated candidate sites calculated as net A-to-G editing (proportion of sequencing reads with alteration of ≥1 A bases to G in treated samples versus untreated samples) in HuH-7 P281L cells that underwent plasmid transfection (top two graphs; n = 2 treated and 2 untreated biological replicates) or LNP treatment (bottom graph; n = 3 treated and 3 untreated biological replicates). Sites with unsuccessful sequencing are omitted. g Off-target editing at PAH1_OT7 site in a dose-response study with HuH-7 P281L cells treated with ABE8.8/PAH1 LNPs (n = 3 biological replicates). Source data are provided as a Source Data file.

Fig. 2. Generation of humanized PKU mice with PAH P281L variant.

a Sanger sequencing chromatograms showing the generation of a humanized mouse model via Cas9-mediated homology-directed repair in mouse zygotes. At the top is the sequence from a wild-type C57BL/6J mouse. At the bottom is the sequence from a mouse homozygous for the humanized Pah P281L allele. The red arrow indicates the site of the P281L variant, and the black arrows indicate the sites of synonymous changes that humanize the local region of the mouse Pah gene. b Age-matched colonymates that are homozygous or heterozygous for the humanized P281L allele. The left picture shows two homozygous mice with PKU as evidenced by hypopigmentation of the fur and two control heterozygous mice with normal fur color, immediately prior to treatment. The right picture shows the two homozygous mice and two heterozygous mice 8 weeks after the homozygous mice received LNP treatment, with normalization of fur color.

Fig. 3. Adenine base editing for correction of PAH P281L variant in humanized PKU mice in vivo.

a Short-term changes in the blood phenylalanine level in homozygous PKU mice (n = 4 animals) following treatment with 2.5 mg/kg dose of ABE8.8/PAH1 LNPs, comparing levels at various time points up to 7 days following treatment to levels in untreated PKU (n = 3 animals) and heterozygous non-PKU control (n = 3 animals) age-matched (8 weeks of age) colonymates (1 blood sample per timepoint). b Short-term changes in the blood phenylalanine level in compound heterozygous PKU mice (n = 2 animals) following treatment with 2.5 mg/kg dose of ABE8.8/PAH1 LNPs, comparing levels at various time points up to 7 days following treatment to levels in untreated heterozygous non-PKU control (n = 2 animals) age-matched (4 weeks of age) colonymates (1 blood sample per timepoint). c Short-term changes in the blood phenylalanine level in homozygous PKU mice (n = 2 animals) and compound heterozygous PKU mice (n = 3 animals) following treatment with 2.5 mg/kg dose of ABE8.8/PAH2 LNPs, comparing levels at various time points up to 7 days following treatment to levels in untreated heterozygous non-PKU control (n = 3 animals) age-matched (8 weeks of age) colonymates (1 blood sample per timepoint). d Long-term changes in the blood phenylalanine level in homozygous PKU mice (n = 3 animals) following treatment with 2.5 mg/kg dose of ABE8.8/PAH1 LNPs, comparing levels at various time points up to 24 weeks following treatment to levels in untreated heterozygous non-PKU control (n = 3 animals) age-matched (8 weeks of age) colonymates (1 blood sample per timepoint). e A-to-G editing in various mouse organs (left, n = 4 animals except for testis, for which n = 2 male animals; right, n = 2 or 3 animals per group), assessed 1–2 weeks following treatment with 2.5 mg/kg dose of LNPs (mean ± standard deviation for each organ). For compound heterozygous mice, each displayed number is % edited P281L alleles (editable alleles) divided by two. Source data are provided as a Source Data file.

Fig. 4. Assessment of mouse liver following LNP treatment.

a Long-term changes in the blood aspartate aminotransferase (AST) level (top) and alanine aminotransferase (ALT) level (bottom) in homozygous PKU mice (n = 3 animals) following treatment with 2.5 mg/kg dose of ABE8.8/PAH1 LNPs, comparing levels at various time points up to 24 weeks following treatment to levels in untreated heterozygous non-PKU control (n = 3 animals) age-matched (8 weeks of age) colonymates (1 blood sample per timepoint). b Liver histology (hematoxylin/eosin staining) at ×20 magnification upon necropsy at 1 week after LNP treatment of humanized PKU mouse (bottom, n = 1 animal) compared to age-matched, untreated non-PKU mouse (top, n = 1 animal). Lines indicate a distance of 50 µm. Source data are provided as a Source Data file.