Mouse model for human arginase deficiency

Abstract

Deficiency of liver arginase (AI) causes hyperargininemia (OMIM 207800), a disorder characterized by progressive mental impairment, growth retardation, and spasticity and punctuated by sometimes fatal episodes of hyperammonemia. We constructed a knockout mouse strain carrying a nonfunctional AI gene by homologous recombination. Arginase AI knockout mice completely lacked liver arginase (AI) activity, exhibited severe symptoms of hyperammonemia, and died between postnatal days 10 and 14. During hyperammonemic crisis, plasma ammonia levels of these mice increased >10-fold compared to those for normal animals. Livers of AI-deficient animals showed hepatocyte abnormalities, including cell swelling and inclusions. Plasma amino acid analysis showed the mean arginine level in knockouts to be approximately fourfold greater than that for the wild type and threefold greater than that for heterozygotes; the mean proline level was approximately one-third and the ornithine level was one-half of the proline and ornithine levels, respectively, for wild-type or heterozygote mice--understandable biochemical consequences of arginase deficiency. Glutamic acid, citrulline, and histidine levels were about 1.5-fold higher than those seen in the phenotypically normal animals. Concentrations of the branched-chain amino acids valine, isoleucine, and leucine were 0.4 to 0.5 times the concentrations seen in phenotypically normal animals. In summary, the AI-deficient mouse duplicates several pathobiological aspects of the human condition and should prove to be a useful model for further study of the disease mechanism(s) and to explore treatment options, such as pharmaceutical administration of sodium phenylbutyrate and/or ornithine and development of gene therapy protocols.

FIG. 1.

Schematic diagram of the AI gene-targeting event. (A) The genomic structure of mouse arginase I (AI) with eight exons and seven introns. (B) The AI gene-replacement vector, containing AI genomic sequences from exon 2 (E2) to exon 8 (E8): exon 4 (E4) is replaced by the neomycin resistance (Neo^R) gene, and the HSV-tk (TK) gene provides strong selection against nonhomologous intergration events. (C) The product of homologous recombination of the AI gene-replacement vector into the mouse genome. The arrows represent the locations of primers used for genotyping the resultant ES cells and mice. 1, primer MAIF/I3; 2, primer NeoF(412); 3, primer MAIR3′UTR(∗+59). PCR A contains primers 1 and 2, and reaction B contains primers 2 and 3. The lines below segment C show the approximate sizes of the different products obtained. See Fig. 2 for an example of results obtained from wild-type, heterozygous, and AI-deficient mice.

FIG. 2.

Genotyping of wild-type, heterozygous, and AI-deficient mice by long-range PCR. Reaction A (lanes 1) contains two AI-specific primers, one located within the targeted region (MAIF/I3) and one outside the targeted region of the AI gene [MAIR3′UTR(∗+59)]. Reaction B (lanes 2) contains a Neo^R gene-specific primer [NeoF(412)] and an AI-specific primer located outside the targeted region [MAIR3′UTR(∗+59)]. For location of the primers on the wild-type and modified AI genes, see Fig. 1. As expected, wild-type mice exhibited a 2.5-kb band in reaction A but no product in reaction B; heterozygous mice exhibited a 2.5-kb and a 4.0-kb (often very weak or absent) band in reaction A and a 3.0-kb band in reaction B; and AI-deficient mice exhibited a strong 4.0-kb band in reaction A and a 3.0-kb band in reaction B.

FIG. 3.

Identification of arginase isoforms in kidney extracts of wild-type (WT), heterozygous (Het), and AI-deficient (KO) mice by immunoprecipitation with an anti-rat AII antibody. Only ∼60% of the arginase activity from kidney extracts of wild-type mice was precipitated by the anti-AII antibody, in agreement with previous data obtained. Almost 100% of arginase activity in AI-deficient and heterozygous mice was precipitated by this antibody, thus proving that arginase AI is not expressed in any organ of the AI knockout mouse. The reason for the lack of arginase AI activity in the kidneys of heterozygous mice is unclear at this time.

FIG. 4.

Histopathologic analysis of AI wild-type and AI-deficient mouse livers (hematoxylin-and-eosin-stained sections) harvested during hyperammonemic episodes. Typical fields (magnification, ×200) from wild-type (A) and knockout (B) mouse livers are shown. AI-deficient mice exhibited a two- to threefold enlargement of hepatocytes, with several different types of inclusions also present (see Results for a detailed description). High magnification (×1,000) views of arginase-deficient mouse (C) and human (D) liver sections are shown. Hepatocytes in the mouse model shared many histological features with those of human patients (see Results for a complete description), including dense, eosinophilic intracytoplasmic inclusion bodies (arrows).

FIG. 5.

Perturbation of amino acid concentrations in plasma of AI-deficient mice in hyperammonemic crisis compared to those of wild-type and heterozygous mice. Normalized mean values obtained from four mice each (wild-type and knockout) or two mice (heterozygotic) were plotted (for raw data, see Results). Compared to wild-type animals, AI-deficient mice exhibited hyperargininemia (approximately fourfold increase), hypoornithinemia (one-half normal levels), and significantly reduced proline levels (one-third wild-type levels). Glutamic acid, histidine (not shown), and citrulline levels were elevated to about 1.5 times those seen in phenotypically normal animals. No significant changes in plasma glutamine concentrations were observed. Concentrations of the branched-chain amino acids (valine, isoleucine, and leucine) were reduced to about 50% of those for phenotypically normal animals, and the aromatics were reduced to ∼60% of those for phenotypically normal animals.