original article © The American Society of Gene & Cell Therapy Human arginase deficiency is characterized by hyper- argininemia and infrequent episodes of hyperammo- nemia that cause neurological impairment and growth retardation. We previously developed a neonatal mouse adeno-associated viral vector (AAV) rh10-mediated therapeutic approach with arginase expressed by a chicken β-actin promoter that controlled plasma ammo- nia and arginine, but hepatic arginase declined rapidly. This study tested a codon-optimized arginase cDNA and compared the chicken β-actin promoter to liver- and muscle-specific promoters. ARG1 −/− mice treated with AAVrh10 carrying the liver-specific promoter also exhibited long-term survival and declining hepatic argi- nase accompanied by the loss of AAV episomes during subsequent liver growth. Although arginase expres- sion in striated muscle was not expected to counteract hyperammonemia, due to muscle’s lack of other urea cycle enzymes, we hypothesized that the postmitotic phenotype in muscle would allow vector genomes to persist, and hence contribute to decreased plasma arginine. As anticipated, ARG1 −/− neonatal mice treated with AAVrh10 carrying a modified creatine kinase-based muscle-specific promoter did not survive longer than controls; however, their plasma arginine levels remained normal when animals were hyperammonemic. These data imply that plasma arginine can be controlled in arginase deficiency by muscle-specific expression, thus suggesting an alternative approach to utilizing the liver for treating hyperargininemia. Received 10 February 2014; accepted 18 May 2014; advance online publication 15 July 2014. doi:10.1038/mt.2014.99 INTRODUCTION Two arginase isoforms are expressed in mammals: arginase 1 (ARG1) and arginase 2 (ARG2). 1 ese are the products of distinct genes located on different chromosomes with independent regula- tion; both are expressed in many cell types, and both are induc- ible by a wide range of agents and in many pathophysiological conditions. 2–4 e arginases catalyze the divalent cation-dependent hydrolysis of l-arginine to form the nonprotein amino acid l-orni- thine and urea. In the liver, in coordination with the other enzymes, this urea cycle reaction constitutes the final step in ureagenesis and is performed by ARG1. 5 ARG1 is expressed most prevalently in hepatocytes and red blood cells, is cytosolic, and is the best char- acterized of the mammalian arginases. Arginase has been detected in a number of nonhepatic tissues that lack a complete urea cycle; in these locations (mainly in the kidney and brain), 5,6 the second isozyme, ARG2, is mitochondrial and the reaction is thought to provide a source of ornithine, the biosynthetic precursor of proline and polyamines. e human type 1 and 2 arginases are related by 58% sequence identity but are immunologically distinct. 6 Arginase deficiency is an autosomal recessive disorder resulting from a loss of ARG1. Neonatal and early infantile presentations are rare and cause severe hyperammonemia 7,8 ; ARG1 deficiency usu- ally presents later in life beginning in late infancy to the second year of life with microcephaly, spasticity, seizures, clonus, loss of ambu- lation (oſten manifesting as spastic diplegia), and failure to thrive associated with hyperargininemia 9 but without profound hyperam- monemia. e neurologic manifestations seen in arginase deficiency may arise from the accumulation of arginine and its metabolites, but recent studies in both humans 10 and mice 11 suggest that guanidino compounds, previously considered the presumptive cause, may not be related to the central nervous system dysfunction characteris- tic of this disorder. e neurologic impairment and developmental Correspondence: Gerald S Lipshutz, 77-120 Center for the Health Sciences, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095-7054, USA. E-mail: [email protected] Myocyte-mediated Arginase Expression Controls Hyperargininemia but not Hyperammonemia in Arginase-deficient Mice Chuhong Hu 1 , Jennifer Kasten 1 , Hana Park 1 , Ragini Bhargava 1 , Denise S Tai 1 , Wayne W Grody 2,3 , Quynh G Nguyen 4 , Stephen D Hauschka 4 , Stephen D Cederbaum 3,5–8 and Gerald S Lipshutz 1,6–11 1 Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 2 Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 3 Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 4 Department of Biochemistry, University of Washington, Seattle, Washington, USA; 5 Department of Human Genetics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 6 Department of Psychiatry, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 7 The Intellectual and Developmental Disabilities Research Center at UCLA, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 8 The Semel Institute for Neuroscience, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 9 Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 10 Department of Urology, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 11 Broad Center of Regenerative Medicine and Stem Cell Research, David Geffen School of Medicine at UCLA, Los Angeles, California, USA 1792 www.moleculartherapy.org vol. 22 no. 10, 1792–1802 oct. 2014
Myocyte-mediated Arginase Expression Controls Hyperargininemia but not Hyperammonemia in Arginase-deficient Mice
Dec 10, 2022
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Myocyte-mediated Arginase Expression Controls Hyperargininemia but not Hyperammonemia in Arginase-deficient Miceoriginal article © The American Society of Gene & Cell Therapy
Human arginase deficiency is characterized by hyper- argininemia and infrequent episodes of hyperammo- nemia that cause neurological impairment and growth retardation. We previously developed a neonatal mouse adeno-associated viral vector (AAV) rh10- mediated therapeutic approach with arginase expressed by a chicken β-actin promoter that controlled plasma ammo- nia and arginine, but hepatic arginase declined rapidly. This study tested a codon-optimized arginase cDNA and compared the chicken β-actin promoter to liver- and muscle-specific promoters. ARG1−/− mice treated with AAVrh10 carrying the liver-specific promoter also exhibited long-term survival and declining hepatic argi- nase accompanied by the loss of AAV episomes during subsequent liver growth. Although arginase expres- sion in striated muscle was not expected to counteract hyperammonemia, due to muscle’s lack of other urea cycle enzymes, we hypothesized that the postmitotic phenotype in muscle would allow vector genomes to persist, and hence contribute to decreased plasma arginine. As anticipated, ARG1−/− neonatal mice treated with AAVrh10 carrying a modified creatine kinase-based muscle-specific promoter did not survive longer than controls; however, their plasma arginine levels remained normal when animals were hyperammonemic. These data imply that plasma arginine can be controlled in arginase deficiency by muscle-specific expression, thus suggesting an alternative approach to utilizing the liver for treating hyperargininemia.
Received 10 February 2014; accepted 18 May 2014; advance online publication 15 July 2014. doi:10.1038/mt.2014.99
INTRODUCTION Two arginase isoforms are expressed in mammals: arginase 1 (ARG1) and arginase 2 (ARG2).1 These are the products of distinct genes located on different chromosomes with independent regula- tion; both are expressed in many cell types, and both are induc- ible by a wide range of agents and in many pathophysiological conditions.2–4 The arginases catalyze the divalent cation-dependent hydrolysis of l-arginine to form the nonprotein amino acid l-orni- thine and urea. In the liver, in coordination with the other enzymes, this urea cycle reaction constitutes the final step in ureagenesis and is performed by ARG1.5 ARG1 is expressed most prevalently in hepatocytes and red blood cells, is cytosolic, and is the best char- acterized of the mammalian arginases. Arginase has been detected in a number of nonhepatic tissues that lack a complete urea cycle; in these locations (mainly in the kidney and brain),5,6 the second isozyme, ARG2, is mitochondrial and the reaction is thought to provide a source of ornithine, the biosynthetic precursor of proline and polyamines. The human type 1 and 2 arginases are related by 58% sequence identity but are immunologically distinct.6
Arginase deficiency is an autosomal recessive disorder resulting from a loss of ARG1. Neonatal and early infantile presentations are rare and cause severe hyperammonemia7,8; ARG1 deficiency usu- ally presents later in life beginning in late infancy to the second year of life with microcephaly, spasticity, seizures, clonus, loss of ambu- lation (often manifesting as spastic diplegia), and failure to thrive associated with hyperargininemia9 but without profound hyperam- monemia. The neurologic manifestations seen in arginase deficiency may arise from the accumulation of arginine and its metabolites, but recent studies in both humans10 and mice11 suggest that guanidino compounds, previously considered the presumptive cause, may not be related to the central nervous system dysfunction characteris- tic of this disorder. The neurologic impairment and developmental
Correspondence: Gerald S Lipshutz, 77-120 Center for the Health Sciences, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095-7054, USA. E-mail: [email protected]
Myocyte-mediated Arginase Expression Controls Hyperargininemia but not Hyperammonemia in Arginase-deficient Mice Chuhong Hu1, Jennifer Kasten1, Hana Park1, Ragini Bhargava1, Denise S Tai1, Wayne W Grody2,3, Quynh G Nguyen4, Stephen D Hauschka4, Stephen D Cederbaum3,5–8 and Gerald S Lipshutz1,6–11
1Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 2Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 4Department of Biochemistry, University of Washington, Seattle, Washington, USA; 5Department of Human Genetics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 6Department of Psychiatry, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 7The Intellectual and Developmental Disabilities Research Center at UCLA, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 8The Semel Institute for Neuroscience, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 9Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 10Department of Urology, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 11Broad Center of Regenerative Medicine and Stem Cell Research, David Geffen School of Medicine at UCLA, Los Angeles, California, USA
1792 www.moleculartherapy.org vol. 22 no. 10, 1792–1802 oct. 2014
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
regression are associated with corticospinal12 and pyramidal tract deterioration. The lack of frequent episodes of hyperammonemia in humans is possibly due to an increase in ARG2 which compensates for the lack of ARG15 (ARG2 has been found to be induced in the kidneys of ARG1-deficient patients)13; if this hypothesis is correct, nonhepatic arginase expression may be an important consideration in developing therapy for these patients. Currently, long-term ther- apy rests on provision of a low-protein diet and administration of sodium benzoate and sodium phenylbutyrate. While these dietary and pharmaceutical interventions can partially alleviate the hyper- argininemia of ARG1 deficiency, no completely effective therapy is available.
We have developed a gene therapy approach for treating ARG1−/− animals with neonatal onset of the disorder by using an adeno-associated viral (AAV) vector expressing ARG114. In these studies, we utilized AAV driven by the ubiquitously express- ing chicken β-actin promoter/cytomegalovirus (CMV) enhancer (CBA) resulting in hepatic arginase expression; however, only very low levels of activity remained after the neonatal period. In these mice, other organs and tissues (particularly the heart and skel- etal muscle) also expressed ARG1, and activity in these remained substantially higher than in liver; whether this had any beneficial effects was unclear. We thus hypothesized that arginase deficiency may be a urea cycle disorder where extrahepatic ARG1 expression
Figure 1 Codon-optimized arginase cDNA improves arginase expression levels. HEK293 cells were transfected with plasmids to assess the function of a codon-optimized version of murine arginase compared to the wild-type cDNA; studies were performed in duplicate. Plasmid DNA was transfected into 293 cells and arginase activity was examined 2 days later.
0
1.4 P = 0.008
Figure 2 Whole-mouse luciferase expression to assess promoter activity and tissue specificity. Mice were injected intravenously with 3 × 1012 genome copies/kg of AAV serotype rh10 expressing luciferase on neonatal day 2. At 3 weeks of age, whole-animal bioluminescent imaging and tissue luminometry was performed to assess expression in different tissues in wild-type animals. (a) In vivo bioluminescent imaging demonstrates photon diffusion patterns among representative images. Left panel: CBA promoter exhibits expression in the heart (red arrow) and the liver (white arrow). Middle panel: Liver-specific thyroxine-binding globulin promoter exhibits expression in the liver (white arrow). Right panel: The striated muscle- specific CK8 regulatory cassette exhibits expression in the heart (red arrow) and skeletal muscle (orange arrow). For all groups, images were acquired with the mice in the ventral position. Images were set with the same references such that side-by-side comparison can be made. After removal of individual tissues, levels of luciferase protein expressed as relative light units (RLU) per μg protein was compared in skeletal muscle (b), the heart (c), and the liver (d). (Data is presented as mean + SD with n of 3–5 per tissue.)
1.0 × 100
1.0 × 101
1.0 × 102
1.0 × 103
1.0 × 104
R LU
/μ g
pr ot
ei n
1.0 × 105
1.0 × 106
1.0 × 101
Molecular Therapy vol. 22 no. 10 oct. 2014 1793
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
in striated muscle might moderate hyperargininemia. The poten- tial therapeutic advantage is that while hepatocytes undergo fre- quent cell division during early life14–16 resulting in a substantial loss of AAV vector genomes, most cardiac and skeletal myocytes are postmitotic at birth17–18 and exhibit greater stability of epi- somal AAV genomes following both adult19 and neonatal vector administration15. This could be advantageous as a neonatal gene therapy approach for hyperargininemia and would have important clinical implications for treating this disorder. In these studies, we have compared the effects of treating ARG1 deficient mice with AAVrh10 vectors carrying a muscle-specific or liver-specific regula- tory cassette driving expression of ARG1 to limit expression to myo- cytes or hepatocytes, respectively.
RESULTS Codon-optimized ARG1 results in increased expression Human 293 cells were transfected in culture by one of three plas- mids to assess change in the level of expression by codon opti- mization (co) of the murine ARG1 cDNA: CBA promoter-ARG1, CBA promoter-ARG1-woodchuck postregulatory enhancer (WPRE), and CBA promoter-coARG1 (Figure 1). The results demonstrated that expression with the codon-optimized form of the cDNA increased expression over 100% of the wild-type cDNA (P = 0.03), while arginase expression was similar between the CBA promoter-ARG1 and CBA promoter-ARG1-WPRE transfected cells (P = 0.70) (overall P value of 0.008).
Comparison of ubiquitous and tissue-specific promoters for expressing ARG1 in hepatic and striated muscle AAVrh10 vectors were developed carrying either the ubiquitously expressing CBA promoter, the liver-specific thyroxine-binding glob- ulin promoter (TBG), or the CK8 regulatory cassette, containing modified components from the mouse M-creatine kinase enhancer and proximal promoter that exhibit high-level striated muscle- specific expression in both cardiomyocytes and skeletal muscle fibers.20 These were ligated to a luciferase transgene and 3 × 1012
genome copies/kg were administered intravenously to mice on the second day of life (DOL) followed by bioluminescence imaging at 3 weeks of age. Figure 2a demonstrates the areas of gene expression in 3-week-old mice. With the TBG promoter, expression is almost exclusively restricted to the liver (white arrow). Animals injected with the CBA promoter exhibit expression throughout most tissues with highest expression in the heart (red arrow) and liver. With the CK8 promoter, bioluminescence is present throughout the animal, due to the bodywide distribution of skeletal muscle,21 but is par- ticularly high in skeletal myocytes of the lower extremities (orange arrow) and abdomen, and the cardiac myocytes of the heart.
Luminometry of dissected tissues corroborated the biolu- minescent findings (Figure 2b–d). At 3 weeks of age, the CBA promoter (white) demonstrated the highest expression in skel- etal muscle (Figure 2b) (1.0 × 105 ± 5.4 × 104 relative light units (RLU)/μg protein), followed by the CK8 promoter (7.6 × 104 ± 2.4 × 104 RLU/μg protein), while the TBG promoter (5.4 × 101 ± 3.7 × 101 RLU/μg protein) had negligible expression (P = 0.51
Figure 3 Rescue of ARG1−/− mice. Survival comparisons in days between untreated ARG1−/− mice (n = 38), littermate controls (n = 34) (heterozygotes), and ARG1−/− mice injected with rAAVrh10-CBA-ARG1-WPRE (n = 45), rAAVrh10-TBG-coARG1 (n = 42), rAAVrh10-CK8-coARG1 (n = 8), rAAVrh10-TBG-luciferase (n = 4), and rAAVrh10-CK8-luciferase (n = 5). (*= end of study).
0 0%
270 360
Figure 4 Arginase activity in different tissues depends on the pro- moter. Total protein was isolated from liver, heart, kidney, and skeletal muscle of either control littermate heterozygotes (a) or ARG1−/− mice that were injected intravenously on neonatal day two with AAV car- rying either the (b) chicken β-actin promoter/CMV enhancer or (c) the liver-specific thyroxine-binding globulin promoter linked to the codon-optimized ARG1 cDNA. Arginase activity was measured with a colorimetric assay determining the quantity of urea converted from arginine by each tissue lysate. Results are expressed as mean ± SD. (Heterozygote, CBA, TBG, respectively: 7 days: n = 23, 9, 12; 1 month: n = 8, 9, 6; 2 months: n = 10, 7, 4; 4 months: n = 4, 4, 4; 8 months: n = 4, 4, 5.)
0 0 50 100 150
Days
Liver Heart Kidney Muscle
Liver Heart Kidney Muscle
Days
Days
1794 www.moleculartherapy.org vol. 22 no. 10 oct. 2014
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
comparing CK8 and CBA promoters, P = 0.01 comparing CK8 and TBG promoters, and P = 0.04 comparing CBA and TBG pro- moters) (Overall P value 0.001). In the heart (Figure 2c), a simi- lar expression pattern was seen: expression was highest with the CBA promoter (6.0 × 105 ± 2.2 × 105 RLU/μg protein), followed by the CK8 promoter (1.6 × 105 ± 1.1 × 105 RLU/μg protein), with minimal expression by the TBG promoter (9.6 × 100 ± 2.5 × 100 RLU/μg protein) (P = 0.04 comparing CBA and CK8 promoters, P = 0.04 comparing CK8 and TBG promoters, and P = 0.01 comparing CBA and TBG promoters) (Overall P < 0.001). Examination of hepatic expression (Figure 2d) demonstrated that the CBA promoter was nearly four times stronger than the TBG promoter (3.8 × 103 ± 3.2 × 103 versus 1.0 × 103 ± 3.6 × 102 RLU/μg protein) whereas expression of the CK8 promoter was only about 10% of the CBA promoter in the liver (3.6 × 102 ± 4.7 × 102 RLU/μg protein) (P = 0.02 comparing CK8 and CBA promoters, P = 0.03 comparing CK8 and TBG promoters, and P = 0.11 comparing CBA and TBG promoters) (Overall P value
0.057). This is probably due to CK8 expression by hepatic myofi- broblasts.21 (All were n = 3–5 per group and data is expressed as average ± SD). We have previously demonstrated that expression does decline, particularly in the liver, over the next several weeks as the animals reach maturity and adult organ size.15 With respect to the relative expression levels of the CBA and CK8 promoters in skeletal and cardiac muscle, it should be noted that CBA drives expression in myocytes as well as in the many nonstriated muscle cells in these tissues, whereas CK8 has almost no transcriptional activity in nonmuscle cells. Thus, the slightly higher luminom- etry data observed with CBA may not be due to higher CBA than CK8 promoter activity in myocytes, but rather due to the addi- tional expression by nonmuscle cells.20
Widespread and liver-only expression of ARG1 results in long-term survival of ARG1-deficient mice ARG1−/− mice received an intravenous injection on the second DOL of one of the following AAV vectors: (i) 3 × 1013 genome
Figure 5 Immunohistochemical analysis of cardiac and skeletal muscle expression of arginase following rAAV-ARG1 administration to ARG1−/− neonatal mice. Neonatal ARG1−/− mice were intravenously injected with AAV-TBG-coARG1, AAV-CK8-coARG1 or AAV-CBA-ARG1. Immunostaining for murine arginase is shown for cardiac muscle: (a) AAV-TBG-coARG1; (b) AAV-CK8-coARG1; (c) AAV-CBA-ARG1) and skeletal muscle: (d) AAV-TBG- coARG1; (e) AAV-CK8-coARG1; (f) AAV-CBA-ARG1). TBG and CBA mice were 4 months old when killed and CK8 mice were 2 weeks old when killed due to their hyperammonemic symptoms. (g) AAV genome copy numbers in liver, skeletal and cardiac muscle were analyzed at selected time points, and are plotted as the mean ± SD (n = 3–5 per group). AAV, adeno-associated viral vector.
a b c
d e f
Molecular Therapy vol. 22 no. 10 oct. 2014 1795
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
copies per kilogram (gc/kg) of rAAVrh10-CBA-ARG1-WPRE (n = 45); (ii) 1 × 1014 gc/kg of rAAVrh10-TBG-coARG1 (n = 42); (iii) 8 × 1013 gc/kg rAAVrh10-CK8-coARG1 (n = 8); (iv) 1 × 1014 gc/kg of rAAVrh10-TBG-luciferase (n = 4); and (v) 1 × 1014 gc/kg
rAAVrh10-CK8-luciferase (n = 5). In addition, a group of litter- mate controls (n = 34) (heterozygotes) and untreated ARG1−/− (n = 38) were included and followed.
In the control groups, DOL two ARG1−/− mice injected intra- venously with 1 × 1014 gc/kg with AAV rh10 carrying a reporter gene (luciferase) driven by either the TBG promoter or the CK8 regulatory cassette all perished by DOL 18 (P < 0.0001 when com- pared with ARG1+/− littermate controls) (Figure 3). Similarly, all untreated ARG1−/− mice were dead by DOL 14 (P < 0.0001 when compared with littermate controls). Untreated ARG1−/− mice and rAAV-luciferase-treated ARG1−/− animals were either found dead or were in extremis and then killed. Other ARG1−/− mice treated with rAAV rh10-ARG1 gene therapy were killed for blood collec- tion and tissue harvesting. The ARG1−/− mice treated with the CBA vector expressing ARG1 exhibited better long-term survival (P = 0.08 compared to littermate controls) than those treated with the TBG vector expressing ARG1 (P = 0.0001 compared to littermate controls), about 90% versus 60% survival to day 240 (end of study for TBG mice), compared to littermate controls (97% survival at 1 year). However, ARG1−/− mice treated with the muscle-specific CK8 promoter expressing ARG1 did not exhibit improved survival with all animals dying by DOL 14: CK8-ARG1 versus CK8-luciferase, P = 0.99; CK8-arginase versus untreated, P = 0.28; CK8-arginase versus normal ARG1+/− littermates, P < 0.0001.
AAV expression in myocytes is long-lived and results in improved control of hyperargininemia compared to hepatocyte-only expression Endogenous ARG1 activity is primarily found in the liver with minimal or no contribution from the heart, kidney, and skeletal muscle (Figure 4a). Hepatic arginase activity also increases sev- eral fold over the first month of life. In contrast, 2-day-old AAV- treated neonatal ARG1−/− mice with AAV expressing arginase from the CBA promoter, exhibit early expression in the liver but this declines over the next 2 months to levels indistinguish- able from the assay background (Figure 4b). As we have previ- ously demonstrated, this decline is due to the rapid loss of AAV episomes in these animals (14 and Figure 5g). However, cardiac expression remains high (with some decline at 8 months of life) and skeletal muscle expression remains modest (Figure 4b). Immunohistochemical examination of tissues in these mice demonstrates myocyte expression of arginase in both cardiac and skeletal muscle (Figure 5c,f). Furthermore and unlike the liver, AAV copy number in cardiac and skeletal myocytes remains relatively stable over this time (Figure 5g). Additionally and irrespective of its nonhepatic source, ARG1, expressed by the CBA promoter provides modest control of ammonia and results in plasma arginine levels being about normal as previ- ously reported (in ref. 14).
A different picture is seen with ARG1−/− mice treated with AAV-expressing arginase from the TBG promoter (Figure 4c) where expression is almost exclusively hepatic. Even though these mice were administered a larger dose with the codon- optimized cDNA to compensate for the decreased strength of the TBG promoter compared to the CBA promoter, early hepatic expression was about half that of the CBA-treated mice (Figure 4b); and then, as seen with the CBA promoter, hepatic
Figure 6 Improvement of plasma urea cycle metabolite levels follow- ing neonatal delivery of rAAV-TBG-coARG1. Plasma metabolite levels were measured at ~1 and 3 weeks, and then monthly thereafter in the AAV-TBG-coARG1-injected ARG1−/− animals. Metabolites examined were: (a) ammonia, (b) arginine, and (c) glutamine. In (a) ammonia levels in untreated ARG1−/− mice were measured in clinically ill animals as these mice all died from hyperammonemia by day 18. Heterozygous controls were included for comparisons. All samples levels are plotted as mean ± SD, n = 3–4 per group. AAV, adeno-associated viral vector.
0 0 30
TBG-coARG1-treated ARG1−/− mice
Day of life
0
500
1,000
1,500
2,000
2,500
Day of life
1796 www.moleculartherapy.org vol. 22 no. 10 oct. 2014
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
expression declined over the first 2 months of life such that the sensitivity of the assay could not detect differences from background.
Ammonia (Figure 6a), arginine (Figure 6b) and gluta- mine (Figure 6c) were less well-controlled in treated AAV- TBG-coARG1-injected ARG1−/− animals compared to control ARG1+/− mice. There…
Human arginase deficiency is characterized by hyper- argininemia and infrequent episodes of hyperammo- nemia that cause neurological impairment and growth retardation. We previously developed a neonatal mouse adeno-associated viral vector (AAV) rh10- mediated therapeutic approach with arginase expressed by a chicken β-actin promoter that controlled plasma ammo- nia and arginine, but hepatic arginase declined rapidly. This study tested a codon-optimized arginase cDNA and compared the chicken β-actin promoter to liver- and muscle-specific promoters. ARG1−/− mice treated with AAVrh10 carrying the liver-specific promoter also exhibited long-term survival and declining hepatic argi- nase accompanied by the loss of AAV episomes during subsequent liver growth. Although arginase expres- sion in striated muscle was not expected to counteract hyperammonemia, due to muscle’s lack of other urea cycle enzymes, we hypothesized that the postmitotic phenotype in muscle would allow vector genomes to persist, and hence contribute to decreased plasma arginine. As anticipated, ARG1−/− neonatal mice treated with AAVrh10 carrying a modified creatine kinase-based muscle-specific promoter did not survive longer than controls; however, their plasma arginine levels remained normal when animals were hyperammonemic. These data imply that plasma arginine can be controlled in arginase deficiency by muscle-specific expression, thus suggesting an alternative approach to utilizing the liver for treating hyperargininemia.
Received 10 February 2014; accepted 18 May 2014; advance online publication 15 July 2014. doi:10.1038/mt.2014.99
INTRODUCTION Two arginase isoforms are expressed in mammals: arginase 1 (ARG1) and arginase 2 (ARG2).1 These are the products of distinct genes located on different chromosomes with independent regula- tion; both are expressed in many cell types, and both are induc- ible by a wide range of agents and in many pathophysiological conditions.2–4 The arginases catalyze the divalent cation-dependent hydrolysis of l-arginine to form the nonprotein amino acid l-orni- thine and urea. In the liver, in coordination with the other enzymes, this urea cycle reaction constitutes the final step in ureagenesis and is performed by ARG1.5 ARG1 is expressed most prevalently in hepatocytes and red blood cells, is cytosolic, and is the best char- acterized of the mammalian arginases. Arginase has been detected in a number of nonhepatic tissues that lack a complete urea cycle; in these locations (mainly in the kidney and brain),5,6 the second isozyme, ARG2, is mitochondrial and the reaction is thought to provide a source of ornithine, the biosynthetic precursor of proline and polyamines. The human type 1 and 2 arginases are related by 58% sequence identity but are immunologically distinct.6
Arginase deficiency is an autosomal recessive disorder resulting from a loss of ARG1. Neonatal and early infantile presentations are rare and cause severe hyperammonemia7,8; ARG1 deficiency usu- ally presents later in life beginning in late infancy to the second year of life with microcephaly, spasticity, seizures, clonus, loss of ambu- lation (often manifesting as spastic diplegia), and failure to thrive associated with hyperargininemia9 but without profound hyperam- monemia. The neurologic manifestations seen in arginase deficiency may arise from the accumulation of arginine and its metabolites, but recent studies in both humans10 and mice11 suggest that guanidino compounds, previously considered the presumptive cause, may not be related to the central nervous system dysfunction characteris- tic of this disorder. The neurologic impairment and developmental
Correspondence: Gerald S Lipshutz, 77-120 Center for the Health Sciences, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095-7054, USA. E-mail: [email protected]
Myocyte-mediated Arginase Expression Controls Hyperargininemia but not Hyperammonemia in Arginase-deficient Mice Chuhong Hu1, Jennifer Kasten1, Hana Park1, Ragini Bhargava1, Denise S Tai1, Wayne W Grody2,3, Quynh G Nguyen4, Stephen D Hauschka4, Stephen D Cederbaum3,5–8 and Gerald S Lipshutz1,6–11
1Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 2Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 4Department of Biochemistry, University of Washington, Seattle, Washington, USA; 5Department of Human Genetics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 6Department of Psychiatry, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 7The Intellectual and Developmental Disabilities Research Center at UCLA, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 8The Semel Institute for Neuroscience, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 9Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 10Department of Urology, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 11Broad Center of Regenerative Medicine and Stem Cell Research, David Geffen School of Medicine at UCLA, Los Angeles, California, USA
1792 www.moleculartherapy.org vol. 22 no. 10, 1792–1802 oct. 2014
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
regression are associated with corticospinal12 and pyramidal tract deterioration. The lack of frequent episodes of hyperammonemia in humans is possibly due to an increase in ARG2 which compensates for the lack of ARG15 (ARG2 has been found to be induced in the kidneys of ARG1-deficient patients)13; if this hypothesis is correct, nonhepatic arginase expression may be an important consideration in developing therapy for these patients. Currently, long-term ther- apy rests on provision of a low-protein diet and administration of sodium benzoate and sodium phenylbutyrate. While these dietary and pharmaceutical interventions can partially alleviate the hyper- argininemia of ARG1 deficiency, no completely effective therapy is available.
We have developed a gene therapy approach for treating ARG1−/− animals with neonatal onset of the disorder by using an adeno-associated viral (AAV) vector expressing ARG114. In these studies, we utilized AAV driven by the ubiquitously express- ing chicken β-actin promoter/cytomegalovirus (CMV) enhancer (CBA) resulting in hepatic arginase expression; however, only very low levels of activity remained after the neonatal period. In these mice, other organs and tissues (particularly the heart and skel- etal muscle) also expressed ARG1, and activity in these remained substantially higher than in liver; whether this had any beneficial effects was unclear. We thus hypothesized that arginase deficiency may be a urea cycle disorder where extrahepatic ARG1 expression
Figure 1 Codon-optimized arginase cDNA improves arginase expression levels. HEK293 cells were transfected with plasmids to assess the function of a codon-optimized version of murine arginase compared to the wild-type cDNA; studies were performed in duplicate. Plasmid DNA was transfected into 293 cells and arginase activity was examined 2 days later.
0
1.4 P = 0.008
Figure 2 Whole-mouse luciferase expression to assess promoter activity and tissue specificity. Mice were injected intravenously with 3 × 1012 genome copies/kg of AAV serotype rh10 expressing luciferase on neonatal day 2. At 3 weeks of age, whole-animal bioluminescent imaging and tissue luminometry was performed to assess expression in different tissues in wild-type animals. (a) In vivo bioluminescent imaging demonstrates photon diffusion patterns among representative images. Left panel: CBA promoter exhibits expression in the heart (red arrow) and the liver (white arrow). Middle panel: Liver-specific thyroxine-binding globulin promoter exhibits expression in the liver (white arrow). Right panel: The striated muscle- specific CK8 regulatory cassette exhibits expression in the heart (red arrow) and skeletal muscle (orange arrow). For all groups, images were acquired with the mice in the ventral position. Images were set with the same references such that side-by-side comparison can be made. After removal of individual tissues, levels of luciferase protein expressed as relative light units (RLU) per μg protein was compared in skeletal muscle (b), the heart (c), and the liver (d). (Data is presented as mean + SD with n of 3–5 per tissue.)
1.0 × 100
1.0 × 101
1.0 × 102
1.0 × 103
1.0 × 104
R LU
/μ g
pr ot
ei n
1.0 × 105
1.0 × 106
1.0 × 101
Molecular Therapy vol. 22 no. 10 oct. 2014 1793
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
in striated muscle might moderate hyperargininemia. The poten- tial therapeutic advantage is that while hepatocytes undergo fre- quent cell division during early life14–16 resulting in a substantial loss of AAV vector genomes, most cardiac and skeletal myocytes are postmitotic at birth17–18 and exhibit greater stability of epi- somal AAV genomes following both adult19 and neonatal vector administration15. This could be advantageous as a neonatal gene therapy approach for hyperargininemia and would have important clinical implications for treating this disorder. In these studies, we have compared the effects of treating ARG1 deficient mice with AAVrh10 vectors carrying a muscle-specific or liver-specific regula- tory cassette driving expression of ARG1 to limit expression to myo- cytes or hepatocytes, respectively.
RESULTS Codon-optimized ARG1 results in increased expression Human 293 cells were transfected in culture by one of three plas- mids to assess change in the level of expression by codon opti- mization (co) of the murine ARG1 cDNA: CBA promoter-ARG1, CBA promoter-ARG1-woodchuck postregulatory enhancer (WPRE), and CBA promoter-coARG1 (Figure 1). The results demonstrated that expression with the codon-optimized form of the cDNA increased expression over 100% of the wild-type cDNA (P = 0.03), while arginase expression was similar between the CBA promoter-ARG1 and CBA promoter-ARG1-WPRE transfected cells (P = 0.70) (overall P value of 0.008).
Comparison of ubiquitous and tissue-specific promoters for expressing ARG1 in hepatic and striated muscle AAVrh10 vectors were developed carrying either the ubiquitously expressing CBA promoter, the liver-specific thyroxine-binding glob- ulin promoter (TBG), or the CK8 regulatory cassette, containing modified components from the mouse M-creatine kinase enhancer and proximal promoter that exhibit high-level striated muscle- specific expression in both cardiomyocytes and skeletal muscle fibers.20 These were ligated to a luciferase transgene and 3 × 1012
genome copies/kg were administered intravenously to mice on the second day of life (DOL) followed by bioluminescence imaging at 3 weeks of age. Figure 2a demonstrates the areas of gene expression in 3-week-old mice. With the TBG promoter, expression is almost exclusively restricted to the liver (white arrow). Animals injected with the CBA promoter exhibit expression throughout most tissues with highest expression in the heart (red arrow) and liver. With the CK8 promoter, bioluminescence is present throughout the animal, due to the bodywide distribution of skeletal muscle,21 but is par- ticularly high in skeletal myocytes of the lower extremities (orange arrow) and abdomen, and the cardiac myocytes of the heart.
Luminometry of dissected tissues corroborated the biolu- minescent findings (Figure 2b–d). At 3 weeks of age, the CBA promoter (white) demonstrated the highest expression in skel- etal muscle (Figure 2b) (1.0 × 105 ± 5.4 × 104 relative light units (RLU)/μg protein), followed by the CK8 promoter (7.6 × 104 ± 2.4 × 104 RLU/μg protein), while the TBG promoter (5.4 × 101 ± 3.7 × 101 RLU/μg protein) had negligible expression (P = 0.51
Figure 3 Rescue of ARG1−/− mice. Survival comparisons in days between untreated ARG1−/− mice (n = 38), littermate controls (n = 34) (heterozygotes), and ARG1−/− mice injected with rAAVrh10-CBA-ARG1-WPRE (n = 45), rAAVrh10-TBG-coARG1 (n = 42), rAAVrh10-CK8-coARG1 (n = 8), rAAVrh10-TBG-luciferase (n = 4), and rAAVrh10-CK8-luciferase (n = 5). (*= end of study).
0 0%
270 360
Figure 4 Arginase activity in different tissues depends on the pro- moter. Total protein was isolated from liver, heart, kidney, and skeletal muscle of either control littermate heterozygotes (a) or ARG1−/− mice that were injected intravenously on neonatal day two with AAV car- rying either the (b) chicken β-actin promoter/CMV enhancer or (c) the liver-specific thyroxine-binding globulin promoter linked to the codon-optimized ARG1 cDNA. Arginase activity was measured with a colorimetric assay determining the quantity of urea converted from arginine by each tissue lysate. Results are expressed as mean ± SD. (Heterozygote, CBA, TBG, respectively: 7 days: n = 23, 9, 12; 1 month: n = 8, 9, 6; 2 months: n = 10, 7, 4; 4 months: n = 4, 4, 4; 8 months: n = 4, 4, 5.)
0 0 50 100 150
Days
Liver Heart Kidney Muscle
Liver Heart Kidney Muscle
Days
Days
1794 www.moleculartherapy.org vol. 22 no. 10 oct. 2014
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
comparing CK8 and CBA promoters, P = 0.01 comparing CK8 and TBG promoters, and P = 0.04 comparing CBA and TBG pro- moters) (Overall P value 0.001). In the heart (Figure 2c), a simi- lar expression pattern was seen: expression was highest with the CBA promoter (6.0 × 105 ± 2.2 × 105 RLU/μg protein), followed by the CK8 promoter (1.6 × 105 ± 1.1 × 105 RLU/μg protein), with minimal expression by the TBG promoter (9.6 × 100 ± 2.5 × 100 RLU/μg protein) (P = 0.04 comparing CBA and CK8 promoters, P = 0.04 comparing CK8 and TBG promoters, and P = 0.01 comparing CBA and TBG promoters) (Overall P < 0.001). Examination of hepatic expression (Figure 2d) demonstrated that the CBA promoter was nearly four times stronger than the TBG promoter (3.8 × 103 ± 3.2 × 103 versus 1.0 × 103 ± 3.6 × 102 RLU/μg protein) whereas expression of the CK8 promoter was only about 10% of the CBA promoter in the liver (3.6 × 102 ± 4.7 × 102 RLU/μg protein) (P = 0.02 comparing CK8 and CBA promoters, P = 0.03 comparing CK8 and TBG promoters, and P = 0.11 comparing CBA and TBG promoters) (Overall P value
0.057). This is probably due to CK8 expression by hepatic myofi- broblasts.21 (All were n = 3–5 per group and data is expressed as average ± SD). We have previously demonstrated that expression does decline, particularly in the liver, over the next several weeks as the animals reach maturity and adult organ size.15 With respect to the relative expression levels of the CBA and CK8 promoters in skeletal and cardiac muscle, it should be noted that CBA drives expression in myocytes as well as in the many nonstriated muscle cells in these tissues, whereas CK8 has almost no transcriptional activity in nonmuscle cells. Thus, the slightly higher luminom- etry data observed with CBA may not be due to higher CBA than CK8 promoter activity in myocytes, but rather due to the addi- tional expression by nonmuscle cells.20
Widespread and liver-only expression of ARG1 results in long-term survival of ARG1-deficient mice ARG1−/− mice received an intravenous injection on the second DOL of one of the following AAV vectors: (i) 3 × 1013 genome
Figure 5 Immunohistochemical analysis of cardiac and skeletal muscle expression of arginase following rAAV-ARG1 administration to ARG1−/− neonatal mice. Neonatal ARG1−/− mice were intravenously injected with AAV-TBG-coARG1, AAV-CK8-coARG1 or AAV-CBA-ARG1. Immunostaining for murine arginase is shown for cardiac muscle: (a) AAV-TBG-coARG1; (b) AAV-CK8-coARG1; (c) AAV-CBA-ARG1) and skeletal muscle: (d) AAV-TBG- coARG1; (e) AAV-CK8-coARG1; (f) AAV-CBA-ARG1). TBG and CBA mice were 4 months old when killed and CK8 mice were 2 weeks old when killed due to their hyperammonemic symptoms. (g) AAV genome copy numbers in liver, skeletal and cardiac muscle were analyzed at selected time points, and are plotted as the mean ± SD (n = 3–5 per group). AAV, adeno-associated viral vector.
a b c
d e f
Molecular Therapy vol. 22 no. 10 oct. 2014 1795
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
copies per kilogram (gc/kg) of rAAVrh10-CBA-ARG1-WPRE (n = 45); (ii) 1 × 1014 gc/kg of rAAVrh10-TBG-coARG1 (n = 42); (iii) 8 × 1013 gc/kg rAAVrh10-CK8-coARG1 (n = 8); (iv) 1 × 1014 gc/kg of rAAVrh10-TBG-luciferase (n = 4); and (v) 1 × 1014 gc/kg
rAAVrh10-CK8-luciferase (n = 5). In addition, a group of litter- mate controls (n = 34) (heterozygotes) and untreated ARG1−/− (n = 38) were included and followed.
In the control groups, DOL two ARG1−/− mice injected intra- venously with 1 × 1014 gc/kg with AAV rh10 carrying a reporter gene (luciferase) driven by either the TBG promoter or the CK8 regulatory cassette all perished by DOL 18 (P < 0.0001 when com- pared with ARG1+/− littermate controls) (Figure 3). Similarly, all untreated ARG1−/− mice were dead by DOL 14 (P < 0.0001 when compared with littermate controls). Untreated ARG1−/− mice and rAAV-luciferase-treated ARG1−/− animals were either found dead or were in extremis and then killed. Other ARG1−/− mice treated with rAAV rh10-ARG1 gene therapy were killed for blood collec- tion and tissue harvesting. The ARG1−/− mice treated with the CBA vector expressing ARG1 exhibited better long-term survival (P = 0.08 compared to littermate controls) than those treated with the TBG vector expressing ARG1 (P = 0.0001 compared to littermate controls), about 90% versus 60% survival to day 240 (end of study for TBG mice), compared to littermate controls (97% survival at 1 year). However, ARG1−/− mice treated with the muscle-specific CK8 promoter expressing ARG1 did not exhibit improved survival with all animals dying by DOL 14: CK8-ARG1 versus CK8-luciferase, P = 0.99; CK8-arginase versus untreated, P = 0.28; CK8-arginase versus normal ARG1+/− littermates, P < 0.0001.
AAV expression in myocytes is long-lived and results in improved control of hyperargininemia compared to hepatocyte-only expression Endogenous ARG1 activity is primarily found in the liver with minimal or no contribution from the heart, kidney, and skeletal muscle (Figure 4a). Hepatic arginase activity also increases sev- eral fold over the first month of life. In contrast, 2-day-old AAV- treated neonatal ARG1−/− mice with AAV expressing arginase from the CBA promoter, exhibit early expression in the liver but this declines over the next 2 months to levels indistinguish- able from the assay background (Figure 4b). As we have previ- ously demonstrated, this decline is due to the rapid loss of AAV episomes in these animals (14 and Figure 5g). However, cardiac expression remains high (with some decline at 8 months of life) and skeletal muscle expression remains modest (Figure 4b). Immunohistochemical examination of tissues in these mice demonstrates myocyte expression of arginase in both cardiac and skeletal muscle (Figure 5c,f). Furthermore and unlike the liver, AAV copy number in cardiac and skeletal myocytes remains relatively stable over this time (Figure 5g). Additionally and irrespective of its nonhepatic source, ARG1, expressed by the CBA promoter provides modest control of ammonia and results in plasma arginine levels being about normal as previ- ously reported (in ref. 14).
A different picture is seen with ARG1−/− mice treated with AAV-expressing arginase from the TBG promoter (Figure 4c) where expression is almost exclusively hepatic. Even though these mice were administered a larger dose with the codon- optimized cDNA to compensate for the decreased strength of the TBG promoter compared to the CBA promoter, early hepatic expression was about half that of the CBA-treated mice (Figure 4b); and then, as seen with the CBA promoter, hepatic
Figure 6 Improvement of plasma urea cycle metabolite levels follow- ing neonatal delivery of rAAV-TBG-coARG1. Plasma metabolite levels were measured at ~1 and 3 weeks, and then monthly thereafter in the AAV-TBG-coARG1-injected ARG1−/− animals. Metabolites examined were: (a) ammonia, (b) arginine, and (c) glutamine. In (a) ammonia levels in untreated ARG1−/− mice were measured in clinically ill animals as these mice all died from hyperammonemia by day 18. Heterozygous controls were included for comparisons. All samples levels are plotted as mean ± SD, n = 3–4 per group. AAV, adeno-associated viral vector.
0 0 30
TBG-coARG1-treated ARG1−/− mice
Day of life
0
500
1,000
1,500
2,000
2,500
Day of life
1796 www.moleculartherapy.org vol. 22 no. 10 oct. 2014
© The American Society of Gene & Cell Therapy Myocyte Arginase Expression for Hyperargininemia
expression declined over the first 2 months of life such that the sensitivity of the assay could not detect differences from background.
Ammonia (Figure 6a), arginine (Figure 6b) and gluta- mine (Figure 6c) were less well-controlled in treated AAV- TBG-coARG1-injected ARG1−/− animals compared to control ARG1+/− mice. There…
Related Documents
