A hypomorphic model of CPS1 deficiency for investigating the effects of hyperammonemia on the developing nervous system

Abstract

Carbamoyl phosphate synthetase 1 (CPS1) deficiency is a rare metabolic disorder that, in neonatal onset, is typically characterized by severe life-threatening and neurologically injuring hyperammonemic episodes with high unmet patient need. Patients that retain limited enzyme activity may present later in life with less severe hyperammonemia. CPS1 drives the first step in the urea cycle, the pathway terrestrial mammals utilize to metabolize nitrogen. In order to probe the effect of hyperammonemia on the developing nervous system and explore new therapies, a murine Cps1 exon 3-4 mutant was previously generated. However, these mice die within 24 h of birth, limiting study capabilities. Herein, we developed a novel Cps1 hypomorphic murine model with residual enzyme activity that maintains survival, but with dysfunction of Cps1 that could be detected biochemically. Characterization, based on the orthologous human variant Asn674Ile, revealed that the variant is reproducible, 100% penetrant and biochemically phenocopies the human disorder. The hypomorph presents with elevated ammonia and glutamate, and reduced citrulline, and with an impaired rate of ureagenesis, providing a novel platform to study and develop therapies for CPS1 deficiency.

Keywords: Carbamoyl phosphate synthetase 1 deficiency; Enzyme activity; Glutamine; Human mutation; Hyperammonemia.

Fig. 1.

The Asn674Ile CPS1 variant is in the bicarbonate phosphorylation region and converts a polar uncharged amino acid to one that is nonpolar and hydrophobic. (A) The domain composition is noted and the amino acid location is identified by the numerals. The mutation is in the C-terminal moiety with bicarbonate phosphorylation, the first two steps in the production of carbamoyl phosphate by phosphorylation of bicarbonate by ATP to produce carboxyphosphate followed by attack by NH₃ to yield carbamate [adapted from de Cima et al. (2015) under the terms of the CC-BY 4.0 license]. (B) Cartoon representation and enlargement (inset) of the ligand-bound human CPS1 protein (PDB ID, 5DOU; de Cima et al., 2015) transparency set to 50%. (Note that colors of the domain correspond with those used in A.) Asn674 is highlighted in red sticks. ADP in the bicarbonate phosphorylation domain is shown in red, orange, yellow and blue sticks. K⁺ and Mg²⁺ ions are shown in purple and green spheres, respectively. Catalytic residues around ADP are shown in light green lines. The yellow dashed line indicates the distance between the center of the Asn674 residue and the center of ADP in angstroms.

Fig. 2.

Cps1 hypomorphic mouse models demonstrate normal liver enzymes while having elevated ammonia and altered response to ammonia loading and carglumic acid administration. (A,B) Values for the liver transaminases alanine aminotransferase and aspartate aminotransferase from plasma of both Cps1^N674I/N674I and Cps1^0/N674I hypomorphic mouse models were similar to wild-type values. (C) Plasma ammonia has a step-wise increase based on the Cps1 mutation, with wild type<Cps1^N674I/N674I<Cps1^0/N674I: 31.00±10.99 µm/l, 46.90±14.35 µm/l and 84.10±35.26 µm/l, respectively. (D) Ammonium challenging with 5 mmol/kg demonstrates a step-wise increase in plasma ammonia based on genotype at 20 min after administration and declines close to baseline at 60 min (n=10 per group). (E) Behavioral testing was performed 15 min after administration. (F) When a higher dose of ammonium is administered (7.5 mmol/kg), greater differences are detected between the hypomorphic and wild-type mice. (G) Behavioral scores were similar between groups reflective of the effect of high blood ammonia levels on the central nervous system. (H) Carglumic acid in drinking water led to a decline in plasma ammonia in wild-type mice but did not in the hypomorphic mice. n=4-5 per group for A,B, n=10 per group for D-F, n=6 per group for G,H, and n=5-6 per group for I. Data presented as mean±s.d. ns, not significant; *P<0.05, **P<0.01 (A-G, one-way ANOVA with Dunnett's T3 multiple comparison test; H, paired two-tailed t-test).

Fig. 3.

Cps1 hypomorphic models demonstrate increased Nags expression and reduced Cps1 protein and enzyme activity, while ureagenesis is maintained. (A) Fold change in hepatic Cps1 RNA expression comparing genotypes was performed normalized to Cps1 wild-type mice (at 1). (B) Fold change in hepatic Nags RNA expression was determined comparing genotypes normalized to Nags wild-type mice (at 1). (C) Representative western blot images from different mouse genotypes examining hepatic Cps1 expression with β-actin loading control (each lane represents a different mouse). (D) Quantitation of western blot of Cps1 protein levels between genotypes demonstrating markedly reduced protein in the hypomorphic mice. (E,F) Representative Cps1 liver immunostaining of the different mouse genotypes (Cps1 in red, glutamine synthetase in green) (E), with average fluorescence intensity per group represented (F). (G) Cps1 enzyme activity was determined from each genotype, showing marked reduction in the hypomorphic mouse livers. (H) [¹⁵N]urea enrichment is decreased in the compound heterozygote, while [¹⁵N]citrulline enrichment is decreased in both the homozygous mutant and the compound heterozygote. [¹⁵N]glutamine is increased in the compound heterozygote. n=10 per group for A,B,D,G; n=8 per group for H; n=3 mice per group for E,F with ROI as n=5 (see Fig. S2). ns, not significant; *P<0.05, ***P<0.001, ****P<0.0001 (A,B,F,G,H, one-way ANOVA with Dunnett's T3 multiple comparison test; D, ordinary one-way ANOVA Tukey's multiple comparison test). AUC, area under the curve; CIT, citrulline; F, female; GLN, glutamine; M, male; WT, wild type. Scale bars: 1 mm.

Fig. 4.

Plasma amino acids show derangements in urea cycle-related amino acids. (A-L) Plasma from adult mice of each genotype was examined for concentration of urea cycle-related amino acid analyte when receiving standard mouse chow with 20% protein (white bars) and when receiving mouse chow with 50% protein (spiculated bars): citrulline (A), ornithine (B), aspartate (C), arginine (D), glutamate (E), glutamine (F), alanine (G), glycine (H), asparagine (I), lysine (J), tyrosine (K) and phenylalanine (L). n=10 per group for standard protein chow; n=6 per group for high-protein chow. Data are presented in µM and as mean±s.d. ns, not significant; *P<0.05, **P<0.01, ***P<0.001 (A-L, one-way ANOVA with Dunnett's T3 multiple comparison test).

Fig. 5.

Hepatic amino acids show some derangements in nitrogen-scavenging amino acids. (A-L) Hepatic lysates from adult mice of each genotype were examined for concentration of each amino acid analyte: citrulline (A), ornithine (B), aspartate (C), arginine (D), glutamate (E), glutamine (F), alanine (G), glycine (H), asparagine (I), lysine (J), tyrosine (K) and phenylalanine (L). n=10 per group). Data are presented in µM and as mean±s.d. ns, not significant; *P<0.05, **P<0.01, ***P<0.001 (A-L, one-way ANOVA with Dunnett's T3 multiple comparison test).

Fig. 6.

Behavioral testing of hypomorphic compound heterozygote receiving standard mouse chow reveals an anxiety-like phenotype. Behavioral phenotype testing was performed on adult mice of each genotype (n=12-17 per group). (A) Novel object recognition testing revealed there was an absence of statistically significant differences between wild-type (black circles), Cps1^N674I/N674I (blue squares) and Cps1^0/N674I (red triangles) mice. (B) In open field testing, measurement of entry to the center was reduced with increasing loss of Cps1. (C) Quantitative measurement of distance traveled was similarly reduced with increasing Cps1 loss. Together, B and C suggest an anxiety-like behavior. (D-F) Light/dark transition testing further suggested anxiety-like behavior: number of entries to the lighted area (D), total amount of time spent in the light side (E) and average speed in the light (F) was reduced for the compound heterozygote. ns, not significant; *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (A-F, one-way ANOVA with Dunnett's T3 multiple comparison test).

Fig. 7.

Conditioned fear testing of hypomorphic compound heterozygote receiving high-protein chow further confirms an anxiety-like phenotype. Behavioral phenotype testing was performed on adult mice of each genotype (n=12-17 per group). In fear conditioning studies of contextual memory in which sexes were combined for analysis, there was evidence of increased freezing in the compound heterozygote. (A) At baseline, freezing was increased in Cps1 compound heterozygote mice. (B) During acquisition, there were no differences during the acquisition (i.e. shock) intertrial intervals (ITIs), demonstrating that mice were freezing following shocks. The only difference was in the final minute of the third ITI as the compound heterozygote demonstrated increased freezing. (C,D) However, upon re-exposure to context 24 h later, total freezing (C) and freezing across the whole session (D) were increased in the compound heterozygote compared to homozygous mutant and wild-type controls, demonstrating that the compound heterozygous mice had stronger contextual fear conditioning memory, consistent with an anxiety-like phenotype. ns, not significant; *P<0.05, **P<0.01 (A,C, ordinary one-way ANOVA with Tukey's multiple comparison test; B,D, two-way ANOVA with mixed effects analysis).

Fig. 8.

CRISPR design showing the endogenous exon 18 of Cps1 and genetic development of Asn674Ile Cps1 hypomorphic mouse model. (A) Pink AAT is the N codon, green C is the silent mutation target to ablate the protospacer adjacent motif (PAM), red cut site shows the cleavage point. The dark green is the gRNA protospacer, light green is the PAM. The sequence of the single-stranded oligodeoxynucleotide repair template shows the 5′ homology arm (blue), 3′ homology arm (red), the introduction of ATT (N>I) and the silent mutation (green). (B) Chromatogram alignment between the identified founder (top) and the wild-type control (bottom). Bottom sequence report shows a 96% contribution of homology-directed repair allele (Ans674Ile) per inference of CRISPR edits (ICE) analysis output. (C) Chromatogram alignment from N1 animals of the potential off-target from chr1:155673955-155673977, showing the region as intact and without off-target indels.