Prevention of hepatocellular adenoma and correction of metabolic abnormalities in murine glycogen storage disease type Ia by gene therapy

Authors

  • Young Mok Lee,

    1. Section on Cellular Differentiation, Program on Developmental Endocrinology and Genetics, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Hyun Sik Jun,

    1. Section on Cellular Differentiation, Program on Developmental Endocrinology and Genetics, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Chi-Jiunn Pan,

    1. Section on Cellular Differentiation, Program on Developmental Endocrinology and Genetics, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Su Ru Lin,

    1. Section on Cellular Differentiation, Program on Developmental Endocrinology and Genetics, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Lane H. Wilson,

    1. Section on Cellular Differentiation, Program on Developmental Endocrinology and Genetics, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Brian C. Mansfield,

    1. Section on Cellular Differentiation, Program on Developmental Endocrinology and Genetics, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    2. Foundation Fighting Blindness, Columbia, MD 21046
    Search for more papers by this author
  • Janice Y. Chou

    Corresponding author
    1. Section on Cellular Differentiation, Program on Developmental Endocrinology and Genetics, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD
    • Building 10, Room 9D42, National Institutes of Health, 10 Center Drive, Bethesda, MD 20892-1830
    Search for more papers by this author
    • fax: 301-402-6035


  • Potential conflict of interest: Nothing to report.

  • See Editorial on Page 1593

Abstract

Glycogen storage disease type Ia (GSD-Ia), which is characterized by impaired glucose homeostasis and chronic risk of hepatocellular adenoma (HCA), is caused by deficiencies in the endoplasmic reticulum (ER)-associated glucose-6-phosphatase-α (G6Pase-α or G6PC) that hydrolyzes glucose-6-phosphate (G6P) to glucose. G6Pase-α activity depends on the G6P transporter (G6PT) that translocates G6P from the cytoplasm into the ER lumen. The functional coupling of G6Pase-α and G6PT maintains interprandial glucose homeostasis. We have shown previously that gene therapy mediated by AAV-GPE, an adeno-associated virus (AAV) vector expressing G6Pase-α directed by the human G6PC promoter/enhancer (GPE), completely normalizes hepatic G6Pase-α deficiency in GSD-Ia (G6pc−/−) mice for at least 24 weeks. However, a recent study showed that within 78 weeks of gene deletion, all mice lacking G6Pase-α in the liver develop HCA. We now show that gene therapy mediated by AAV-GPE maintains efficacy for at least 70-90 weeks for mice expressing more than 3% of wild-type hepatic G6Pase-α activity. The treated mice displayed normal hepatic fat storage, had normal blood metabolite and glucose tolerance profiles, had reduced fasting blood insulin levels, maintained normoglycemia over a 24-hour fast, and had no evidence of hepatic abnormalities. After a 24-hour fast, hepatic G6PT messenger RNA levels in G6pc−/− mice receiving gene therapy were markedly increased. Because G6PT transport is the rate-limiting step in microsomal G6P metabolism, this may explain why the treated G6pc−/− mice could sustain prolonged fasts. The low fasting blood insulin levels and lack of hepatic steatosis may explain the absence of HCA. Conclusion: These results confirm that AAV-GPE–mediated gene transfer corrects hepatic G6Pase-α deficiency in murine GSD-Ia and prevents chronic HCA formation. (HEPATOLOGY 2012;56:1719–1729)

Glycogen storage disease type Ia (GSD-Ia or von Gierke disease) is caused by a deficiency in glucose-6-phosphatase-α (G6Pase-α or G6PC), an enzyme that is expressed primarily in the liver, kidney, and intestine.1 G6Pase-α catalyzes the hydrolysis of glucose-6-phosphate (G6P) to glucose and phosphate in the terminal step of glycogenolysis and gluconeogenesis. Patients affected by GSD-Ia are unable to maintain glucose homeostasis and present with fasting hypoglycemia, growth retardation, hepatomegaly, nephromegaly, hyperlipidemia, hyperuricemia, and lactic acidemia.1 There is no cure for GSD-Ia. Hypoglycemia can be managed using dietary therapies2, 3 that enable patients to attain near normal growth and pubertal development. However, the longer term clinical complications, and their underlying pathological processes, remain uncorrected. One of the most significant chronic risks is hepatocellular adenoma (HCA), which develops in 70%-80% of GSD-Ia patients >25 years of age.1, 4, 5 HCAs in GSD-Ia patients are small, multiple, and nonencapsulated, with complications including local compression and intratumoral hemorrhage. In 10% of GSD-Ia patients, HCAs undergo malignant transformation to hepatocellular carcinoma (HCC).1, 5, 6

G6Pase-α is a hydrophobic protein anchored in the endoplasmic reticulum (ER) by nine transmembrane helices.1 Gene therapy studies using adeno-associated virus (AAV) vectors carrying G6Pase-α have been performed in animal models of GSD-Ia and, in short-term studies, demonstrated efficacy in the absence of toxicity (reviewed by Chou and Mansfield7). We have shown that systemic administration of AAV-GPE, an AAV pseudotype 2/8 vector containing the human G6Pase-α coding sequence along with ≈3 kb of the G6PC gene 5′-flanking promoter/enhancer region (GPE) delivered the G6Pase-α transgene to the liver of G6Pase-α-null (G6pc−/−) mice and restored wild-type (WT) levels of hepatic G6Pase-α activity for the entire duration of a 24-week study.8 The infused G6pc−/− mice exhibited normal blood glucose and metabolite profiles, normal hepatic glycogen and fat storage, and the absence of any apparent immune response. This raised the possibility that if chronic complications of GSD-Ia were also resolved, the AAV-GPE therapy might be a viable clinical option. The AAV vectors have low immunogenicity, exhibit few toxic effects, and are not known to be associated with any human disease.7, 9, 10 However, in addition to the chronic risk of HCA and HCC in GSD-Ia, Donsante et al.11 have also shown an increased incidence of HCC associated with the use of an AAV vector in mice. In this long-term gene therapy study, we examine this combined vector and disease risk for hepatic neoplasia.

Although G6pc−/− mice exhibit a clinical phenotype that mimics the symptoms of human GSD-Ia,12 they have a short life span that complicates studies on the chronic risk of HCA development. Recently, Mutel et al.13 generated liver-specific G6pc-null mice that exhibited a markedly improved survival rate compared with the global knockout mice. They showed that the aged liver-specific G6pc-null mice developed hepatomegaly and marked hepatic steatosis. The first detectable HCA occurred 36 weeks after gene deletion, with all mice developing HCA within 78 weeks of gene deletion.13 In this study, we examined the safety and efficacy of gene therapy mediated by AAV-GPE in murine GSD-Ia throughout a 90-week period and found that this vector not only corrected metabolic abnormalities in murine GSD-Ia but also showed no evidence of HCA formation.

Abbreviations

AAV, adeno-associated virus; COX-2, cyclooxygenase-2; ER, endoplasmic reticulum; FBPase-1, fructose-1,6-bisphosphatase; G6P, glucose-6-phosphate; G6Pase-α, glucose-6-phosphatase-α; G6PDH, G6P dehydrogenase; G6PT, glucose-6-phosphate transporter; GPE, human G6PC promoter and enhancer; qPCR, quantitative polymerase chain reaction; GSD-Ia, glycogen storage disease type Ia; HCA, hepatocellular adenoma; HCC, hepatocellular carcinoma; mRNA, messenger RNA; PFK-1, phosphofructokianse-1; RT-PCR, reverse-transcription polymerase chain reaction; SREBP-1c, sterol regulatory element binding protein-1c; vp, viral particles; WT, wild-type.

Materials and Methods

Infusion of G6pc−/− Mice With AAV-GPE.

All animal studies were conducted using an animal protocol approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Animal Care and Use Committee. The AAV-GPE vector8 was infused into G6pc−/− mice12 via the retro-orbital sinus. Age-matched G6pc+/+/G6pc+/− as well as 6- to 10-week-old G6pc−/− mice were used as controls. For the virus-infused mice, glucose therapy12 was terminated immediately after infusion.

Glucose tolerance testing of mice consisted of fasting for 6 hours, prior to blood sampling, followed by intraperitoneal injection of a glucose solution at 2 mg/g body weight, and repeated blood sampling via the tail vein for 2 hours.

Phosphohydrolase and Microsomal G6P Uptake Assays.

Microsome isolation, phosphohydrolase assays, enzyme histochemical analysis of G6Pase-α, and microsomal G6P uptake assays were performed as described.8, 12

Quantification of Vector DNA.

Total DNA was isolated from liver tissues using the GenElute Mammalian Genomic DNA Miniprep Kits (Sigma-Aldrich, St. Louis, MO). The vector genome numbers were quantified via real-time quantitative polymerase chain reaction (qPCR) using the Applied Biosystems 7300 Real-Time PCR System (Foster City, CA). The following TaqMan probes were used: human G6PC gene, Hs00609178_m1; mouse β-actin, Mm00607939_s1. Data were analyzed using SDS version 1.3 software (Applied Biosystems) and were normalized to β-actin DNA. Plasmid DNA corresponding to 0.01 to 100 copies of human G6PC gene was used in a standard curve. To determine the vector genome copy number, the Ct values of sample were compared with the standard curve.

Phenotype Analyses.

Mice were first examined for hepatic nodules via ultrasound using the Vevo 2100 system (VisualSonics, Toronto, Canada), and blood samples were collected from the tail vein. Blood glucose, total cholesterol, and uric acid were analyzed using kits obtained from Thermo Electron (Louisville, CO); triglycerides, using a kit from Sigma Diagnostics (St. Louis, MO); lactate, using a kit from Trinity Biotech (St. Louis, MO); and insulin, using an ultrasensitive mouse insulin enzyme-linked immunosorbent assay kit from Crystal Chem (Downers Grove, IL). Hepatic glycogen contents were measured as described.8 To determine hepatic triglyceride, glucose, and G6P contents, liver tissues were homogenized in radio immunoprecipitation assay buffer (50 mM Tris HCl [pH 8.0], 150 mM NaCl, 1% Triton X-100, 0.5% Na-deoxycholate, and 0.1% sodium dodecyl sulfate; Thermo Scientific, Rockford, IL). Triglycerides were measured using a kit from Sigma Diagnostics; glucose, using a kit from Thermo Electron; and G6P, using a kit from BioVision (Mountain View, CA).

For hematoxylin and eosin and oil red O staining,8 liver sections were preserved in 10% neutral buffered formalin and sectioned at 4-10 microns thickness. The stained sections were visualized using the Axioskop2 plus microscope and the AxioVision 4.5 software (Carl Zeiss, Thornwood, NY).

Quantitative Real-Time RT-PCR and Antibody Analysis.

Total RNAs were isolated from liver tissues using the TRIzol Reagent (Invitrogen, Carlsbad, CA). The messenger RNA (mRNA) expression was quantified via real-time reverse transcription (RT)-PCR using the Applied Biosystems 7300 Real-Time PCR System with TaqMan probes (Foster City, CA). Data were analyzed using SDS version 1.3 software and normalized to β-actin RNA. Antibodies against human G6Pase-α were detected via western blot analysis as described.8

Statistical Analysis.

An unpaired t test was performed using the GraphPad Prism Program, version 4 (San Diego, CA). Values were considered statistically significant at P < 0.05.

Results

AAV-GPE Infusion Directs Long-Term Hepatic G6Pase-α Expression.

We infused 2- or 4-week-old G6pc−/− mice (n = 18) with varying doses of AAV-GPE (5 × 1012 to 3 × 1013 viral particles [vp]/kg) predicted to restore and maintain 3%-100% of WT hepatic G6Pase-α activity. We also infused one 15-week-old (5 × 1012 vp/kg) and one 30-week-old (1 × 1013 vp/kg) G6pc−/− mouse. The low survival rate of GSD-Ia mice under glucose therapy severely restricted the numbers of adult mice available to study.12 Metabolic and histological profiles of the 20 infused animals were monitored for 70-90 weeks, and all measurements were compared with those of their G6pc+/+ and G6pc+/− littermates. The phenotype of both G6pc+/+ and G6pc+/− mice are indistinguishable from WT.12

There were no premature deaths of G6pc−/− mice receiving gene therapy. Hepatic G6Pase-α activity and glycogen content were assessed in mice sacrificed after a 24-hour fast. The mean fasting hepatic G6Pase-α activity of 70- to 90-week-old WT mice (n = 20) was 185.8 ± 12.7 nmol/mg/minute (Fig. 1A). As planned, there was a range of hepatic G6Pase-α activities restored in the treated mice. Of the 20 AAV-GPE–treated G6pc−/− mice sacrificed at age 70-90 weeks, six mice had low levels (3%-9% of WT activity) of hepatic G6Pase-α activity and were designated AAV-L; nine mice had medium levels (22%-63% of WT activity) of hepatic G6Pase-α activity and were designated AAV-M; and five mice had high levels (81%-128% of WT activity) of hepatic G6Pase-α activity and were designated AAV-H (Fig. 1A). Real-time RT-PCR analysis showed a linear relationship between hepatic G6Pase-α mRNA expression and G6Pase-α activity (Fig. 1B). Likewise, hepatic G6Pase-α activity also increased with hepatic vector genome copy numbers (Fig. 1C).

Figure 1.

Hepatic G6Pase-α activity and mRNA expression in 70- to 90-week-old WT and AAV-GPE–treated G6pc−/− mice following a 24-hour fast. G6pc−/− mice at age 2 weeks (n = 7), 4 weeks (n = 11), 15 weeks (n = 1 [*]), and 30 weeks (n = 1 [**]) were studied. (A) Hepatic G6Pase-α activity is shown at the indicated ages in weeks (W). The mice are grouped based on their G6Pase-α activity relative to WT activity as low (AAV-L), medium (AAV-M), and high (AAV-H). (B) Hepatic G6Pase-α mRNA expression and its relationship to G6Pase-α activity in treated G6pc−/− mice. Data are presented as the mean ± SEM. *P < 0.05. **P < 0.005. (C) Relationship between hepatic G6Pase-α mRNA expression and vector genome copy numbers in treated G6pc−/− mice. (D) Histochemical analysis of hepatic G6Pase-α activity. Freshly sectioned liver specimens were analyzed for G6Pase-α activity using the method of lead trapping of phosphate generated by G6P hydrolysis.8 Each image represents an individual mouse, so two mice are shown for each treatment. Treatments are indicated by the following notation: −/−, untreated G6pc−/− mice; +/+, WT mice; −/− AAV, G6pc−/− mice infused with various dosages of AAV-GPE. AAV-L (n = 6), AAV-M (n = 9), and AAV-H (n = 5) are AAV-GPE–treated G6pc−/− mice expressing 3%-9%, 22%-63%, and 81%-128% normal hepatic G6Pase-α activity, respectively.

Enzyme histochemical analysis showed that G6Pase-α in WT mice was distributed throughout the liver, with significantly higher levels in proximity to blood vessels. There was no stainable G6Pase-α activity in the liver sections of untreated G6pc−/− mice (Fig. 1D). In AAV-GPE–treated G6pc−/− mice, G6Pase-α was also distributed throughout the liver but with foci, not related to blood vessels, containing markedly higher levels of enzymatic activity (Fig. 1D). The uneven distribution of hepatic G6Pase-α in the treated G6pc−/− mice suggests that a substantial proportion of hepatocytes harbored low or little G6Pase-α, including AAV-H livers expressing 81%-128% of WT G6Pase-α activity. Uniform hepatic G6Pase-α expression is not required for rescue of the GSD-Ia phenotype.

AAV-GPE Infusion Corrects Metabolic Abnormalities in GSD-Ia.

GSD-Ia is characterized by hypoglycemia, hypercholesterolemia, hypertriglyceridemia, hyperuricemia, and lactic acidemia.1 Blood glucose levels in AAV-H mice expressing WT hepatic G6Pase-α activity were indistinguishable from those of the control littermates (Fig. 2A). AAV-M and AAV-L expressing 22%-63% and 3%-9% normal hepatic G6Pase-α activity, respectively, also maintained euglycemia (≈100 mg/dL) state,14 but their blood glucose levels were consistently lower than the control littermates (Fig. 2A). All treated G6pc−/− mice exhibited normal serum profiles of cholesterol and triglyceride, whereas serum levels of uric acid and lactic acid in the treated G6pc−/− mice were lower than those in the control littermates (Fig. 2A).

Figure 2.

Phenotype analysis of AAV-GPE–treated G6pc−/− mice at age 70-90 weeks. (A) Blood glucose, triglyceride, cholesterol, uric acid, and lactic acid levels. (B) Body weight, body length, and body mass index. F, females; M, males. (C) Liver weight. Treatments are indicated by the following notation: +/+, WT mice; −/−AAV, G6pc−/− mice infused with various dosages of AAV-GPE. AAV-L (n = 6), AAV-M (n = 9), and AAV-H (n = 5) are AAV-GPE-treated G6pc−/− mice expressing 3%-9%, 22%-63%, and 81%-128% normal hepatic G6Pase-α activity, respectively. Data are presented as the mean ± SEM. *P < 0.05. **P < 0.005.

The average body weights of female and male AAV-GPE–treated mice at age 70-90 weeks were 70% and 62%, respectively, of their age- and sex-matched control mice (Fig. 2B). However, the average body lengths of the treated G6pc−/− mice were 90% of the controls (Fig. 2B). Consequently, body mass index values15 of the treated G6pc−/− mice were significantly lower than those of the control littermates (Fig. 2B). Although body mass index values of both mouse groups signify normal growth,15 the treated G6pc−/− mice were considerably leaner. The liver weights in WT mice were relatively constant (Fig. 2C). In AAV-GPE–infused mice, liver weights were inversely correlated to the hepatic G6Pase-α activity restored (Fig. 2C). When liver weights were expressed as percent of body weight, the treated G6pc−/− mice had significantly higher values because of their lower body weights. However, when absolute liver weights were compared directly, there was no significant difference between AAV-M, AAV-H, and control littermates (Fig. 2C). AAV-L mice, however, continued exhibiting hepatomegaly. AAV-GPE delivers little or no transgene to the kidney.8 However, the infused mice expressing higher hepatic G6Pase-α activity had lower kidney weights (data not shown), suggesting good hepatic metabolic control normalized nephromegaly.

Absence of Histological Abnormalities, Steatosis, or HCA in AAV-GPE–Infused G6pc−/− Livers.

To determine the presence of HCA nodules in AAV-GPE–treated G6pc−/− mice, we first conducted ultrasound analysis, followed by extensive examination of the livers and histological analysis of liver biopsy samples, using five or more separate sections per liver. Ultrasound and morphological analyses detected no hepatic nodules in WT (n = 20) and the transduced G6pc−/− (n = 20) mice that lived to age 70-90 weeks. The AAV-GPE–treated G6pc−/− mice infused at age 2 or 4 weeks (n = 18) exhibited no hepatic histological abnormalities (Fig. 3A) except increased glycogen storage (Fig. 3A,B). The 84-week-old mouse infused at age 15 weeks, which expressed 6% of normal hepatic G6Pase-α activity, exhibited elevated glycogen storage and a few necrotic foci in one liver section (Fig. 3C). Although most of the liver tissue sections of the 90-week-old mouse infused at age 30 weeks, which expressed 38% of normal hepatic G6Pase-α activity, exhibited no histological abnormalities, one liver section did have many necrotic foci (Fig. 3C). Because necrotic foci are a characteristic hepatic pathology seen in untreated GSD-Ia mice age 6 weeks or older,16 it is quite likely that the necrotic foci had developed before initiation of gene therapy at age 15 or 30 weeks.

Figure 3.

Histological, glycogen, lipid, and COX-2 analyses in the livers of 24-hour fasted WT and AAV-GPE–treated G6pc−/− mice at age 70-90 weeks. (A) Hematoxylin and eosin–stained liver sections from treated G6pc−/− mice infused at age 2 or 4 weeks (magnification ×200). Each plate represents an individual mouse, so two mice are shown for each treatment. (B) Hepatic glycogen contents. (C) Hematoxylin and eosin–stained liver sections from AAV-GPE–treated G6pc−/− mice infused at age 15 or 30 weeks (magnification ×200). Duplicated plates are shown for each mouse. (D) Hepatic triglyceride contents. (E) Oil red O staining (magnification ×400). Each image represents an individual mouse, so two mice are shown for each treatment. (F) Quantification of COX-2 mRNA by real-time RT-PCR. Treatments are indicated by the following notation: −/−, untreated G6pc−/− mice; +/+, WT mice (n = 20); −/−AAV, AAV-GPE–treated G6pc−/− mice (n = 20). AAV-L (n = 6), AAV-M (n = 9), AAV-H (n = 5) are the treated G6pc−/− mice expressing 3%-9%, 22%-63%, and 81%-128% normal hepatic G6Pase-α activity, respectively. Data are presented as the mean ± SEM. *P < 0.05. **P < 0.005.

The livers of liver-specific G6pc-null mice were reported to develop HCA with marked steatosis.13 Although a few WT mice had increased hepatic fat storage, there was little or no fat storage in the livers of AAV-GPE–treated G6pc−/− mice (Fig. 3A). Moreover, hepatic triglyceride contents in the treated G6pc−/− mice (n = 20) were not statistically different from those in WT mice (Fig. 3D). Oil red O staining confirmed that lipid contents in the treated animals (n = 20) were similar to that in the controls (Fig. 3E).

It has been well established that cyclooxygenase-2 (COX-2) is a marker that is over-expressed in many pre-malignant and malignant cancers, including HCC.17 Quantitative RT-PCR analysis showed that similar levels of hepatic COX-2 message were expressed in AAV-GPE-treated G6pc−/− and control littermates (Fig. 3F).

AAV-GPE–Treated G6pc−/− Mice Exhibit Normal Fasting Glucose and Glucose Tolerance Profiles.

The mean blood glucose levels of WT mice (n = 20) before commencing fasting were 165.0 ± 3.0 mg/dL (zero time) which decreased to 113.3 ± 6.5 mg/dL after 24 hours of fasting (Fig. 4A). The fasting blood glucose profiles of AAV-L and AAV-M mice paralleled those of the control mice, but blood glucose levels were consistently lower (Fig. 4A), while the fasting glucose profile of AAV-H mice was indistinguishable from that of the control mice. In sharp contrast, untreated G6pc−/− mice, exhibited marked hypoglycemia within 60-75 minutes of fasting (Fig. 4B), a hallmark of GSD-Ia.1 In summary, AAV-GPE-treated G6pc−/− mice no longer suffered from the fasting hypoglycemia characteristic of GSD-Ia.1

Figure 4.

Fasting blood glucose and glucose tolerance profiles. (A) Fasting blood glucose profiles in WT and AAV-GPE–treated G6pc−/− mice at age 70-90 weeks. (B) Fasting blood glucose profiles in untreated G6pc−/− mice at age 6-8 weeks. (C) Glucose tolerance profiles in WT and treated G6pc−/− mice at age 70-90 weeks. WT or treated G6pc−/− mice were fasted for 6 hours, injected intraperitoneally with 2 mg/g dextrose, and sampled for blood every 30 minutes via the tail vein. Data are presented as the mean ± SEM. Treatments are indicated by the following notation: +/+, WT mice; −/−, untreated G6pc−/− mice. AAV-L (n = 6), AAV-M (n = 9), and AAV-H (n = 5) are AAV-GPE–treated G6pc−/− mice expressing 3%-9%, 22%-63%, and 81%-128% normal hepatic G6Pase-α activity, respectively.

Blood glucose tolerance profiles in AAV-M and AAV-H mice were indistinguishable from those of WT littermates (Fig. 4C). In AAV-L mice, following intraperitoneal glucose injection, blood glucose levels declined at a faster rate than the WT controls.

Reduced Fasting Blood Insulin Levels in AAV-GPE–Treated G6pc−/− Mice.

Insulin signaling regulates hepatic glucose and lipid metabolism.18 After 24 hours of fasting, blood insulin levels in 70- to 90-week-old WT (n = 20) and treated G6pc−/− mice (n = 20) were 1.84 ± 0.29 and 0.56 ± 0.09 ng/mL, respectively (Fig. 5A). Both were within the normal range,19 although fasting blood insulin levels in the AAV-GPE–treated G6pc−/− mice were more close to the normal average values.19 While fasting blood insulin levels in the treated G6pc−/− mice did not correlate with hepatic G6Pase-α restored, insulin levels in WT and the treated G6pc−/− mice exhibited a linear relationship to their body weights (Fig. 5A).

Figure 5.

Blood insulin and hepatic mRNA levels for SREBP-1c and glucokinase in 70- to 90-week-old WT and AAV-GPE–treated G6pc−/− mice after 24 hours of fast. (A) Fasting blood insulin levels and their relationship to animal body weights. (B) Quantification of SREBP-1c mRNA via real-time RT-PCR. (C) Quantification of glucokinase mRNA and the relationship of fasting blood insulin to hepatic glucokinase mRNA levels. Treatments are indicated by the following notation: +/+ (○), WT mice (n = 20); −/−AAV (•), AAV-GPE–treated G6pc−/− mice (n = 20). Data are presented as the mean ± SEM. **P< 0.005.

The transcriptional effect of insulin is mediated by sterol regulatory element binding protein-1c (SREBP-1c).18 Quantitative RT-PCR analysis showed that similar levels of hepatic SREBP-1c transcripts were expressed in the 24-hour-fasted AAV-GPE–treated G6pc−/− and control mice (Fig. 5B). Glucokinase is a glucose sensor.20 Hepatic glucokinase activity decreases when blood insulin levels are low, as when fasting.20 As was seen with blood insulin, after 24 hours of fast, hepatic glucokinase transcripts in the treated G6pc−/− mice were significantly lower than that in the control littermates (Fig. 5C). Interestingly, levels of hepatic glucokinase mRNA and fasting blood insulin exhibited a linear relationship in both WT and AAV-GPE–treated G6pc−/− mice (Fig. 5C).

Glucose Homeostasis in the Livers of AAV-GPE–Infused G6pc−/− Mice.

During fasting, blood glucose homeostasis is maintained by endogenous glucose produced in the liver from hydrolysis of G6P by the glucose-6-phosphate transporter (G6PT)/G6Pase-α complex in the terminal step of gluconeogenesis and glycogenolysis.1 G6pc−/− mice, lacking a functional G6Pase-α, are incapable of producing endogenous glucose in the liver, kidney, or intestine. After 24 hours of fast, hepatic free glucose levels in WT mice (n = 20) were 389 ± 17 nmol/mg protein (Fig. 6A) and in AAV-L (n = 6), AAV-M (n = 9), and AAV-H (n = 5) mice were 61%, 68%, and 90%, respectively, of that in WT mice (Fig. 6A). Intracellular G6P levels in the fasted AAV-L and AAV-M livers were 2.9-fold and 1.6-fold, respectively, higher than WT livers, but intracellular G6P levels in the fasted AAV-H livers were statistically similar to that of WT livers (Fig. 6A).

Figure 6.

Glucose homeostasis and analysis of anti–G6Pase-α antibody in WT and AAV-GPE–treated G6pc−/− mice. Hepatic levels of free glucose, G6P, mRNA for PEPCK-C, FBPase-1, PGMase, G6PT, PFK-1, and G6PDH were determined in the livers of WT and the treated G6pc−/− mice after a 24-hour fast at age 70-90 weeks. (A) Hepatic free glucose and G6P levels. (B, C) Quantification of PEPCK-C, FBPase-1, PGMase, G6PT (B), PFK-1, and G6PDH (C) mRNA levels by real-time RT-PCR. (D) Microsomal G6P uptake activity. Data are presented as the mean ± SEM. *P< 0.05. **P< 0.005. (E) Analysis of anti–G6Pase-α antibody in the sera of 70- to 90-week-old WT and AAV-GPE–treated G6pc−/− mice. Microsomal proteins from Ad-human G6Pase-α infected COS-1 cells were electrophoresed through a single 12% polyacrylamide–sodium dodecyl sulfate gel and transferred onto a PVDF membrane.8 Membrane strips, representing individual lanes on the gel were individually incubated with the appropriate serum. A monoclonal antibody against human G6Pase-α that also recognizes murine G6Pase-α8 was used as a positive control (lane 1). Treatments are indicated by the following notation: +/+, WT mice; −/−AAV, G6pc−/− mice infused with various dosages of AAV-GPE. AAV-L (n = 6), AAV-M (n = 9), and AAV-H (n = 5) are the treated G6pc−/− mice expressing 3%-9%, 22%-63%, and 81%-128% normal hepatic G6Pase-α activity, respectively.

Hepatic G6P participates in several metabolic pathways, including glycogen synthesis, glycolysis, the pentose-phosphate pathway, in the cytoplasm, and endogenous glucose production in the ER lumen (Fig. 7). We examined hepatic expression of several key enzymes involved in the aforementioned pathways in mice after a 24-hour fast. These included: the cytosolic phosphoenolpyruvate carboxykinase (PEPCK-C) that catalyzes the first committed step in hepatic gluconeogenesis;21 fructose-1,6-bisphosphatase (FBPase-1) that converts fructose-1,6-bisphosphate to fructose-6-phosphate;22 phosphoglucomutase (PGMase) that catalyzes the reversible conversion of G6P and glucose-1-P in glycogenolysis and glycogen synthesis;22 phosphofructokianse-1 (PFK-1) that catalyzes the irreversible rate-limiting step in glycolysis by converting fructose-6-P to fructose-1,6-diphosphate;22 G6P dehydrogenase (G6PDH) that catalyzes the first reaction in the pentose-phosphate pathway by converting G6P to 6-phosphgluconolactone;23 and G6PT that transports cytoplasmic G6P into the ER lumen1 (Fig. 7). Quantitative real-time RT-PCR analysis showed that in the fasted livers, PEPCK-C and PGMase transcripts were unchanged, while FBPase-1 transcripts were increased in the AAV-GPE–treated G6pc−/− mice compared with controls (Fig. 6B). Both PFK-1 and G6PDH transcripts in AAL-L livers were increased although in AAV-M and AAV-H livers they were still statistically similar to that of WT livers (Fig. 6C). Interestingly, hepatic G6PT mRNA levels in the treated G6pc−/− mice were 2.2-fold higher than the WT controls (Fig. 6B), regardless of the levels of hepatic G6Pase-α activity restored. The G6PT-mediated hepatic microsomal G6P uptake activity is the rate-limiting in endogenous glucose production,24 but is codependent upon G6Pase-α activity.12 Hepatic microsomes prepared from G6pc−/− mice, with an intact G6PT, exhibit markedly lower G6P uptake activity compared with WT hepatic microsomes,12 which can be reversed if G6Pase-α activity is restored via gene transfer.25 In AAV-L, AAV-M, and AAV-H livers, microsomal G6P uptake activities were 43%, 50%, and 72%, respectively, of WT activity (Fig. 6D), reflecting the increase in hepatic G6Pase-α activity (Fig. 1A) that paralleled hepatic free glucose levels (Fig. 6A).

Figure 7.

Pathways for G6P metabolism in normal and GSD-Ia liver during fasting. During fasting, G6P, the end product of gluconeogenesis and glycogenolysis, is transported from the cytoplasm into the lumen of the ER by G6PT.1 Inside the ER, G6P is hydrolyzed by G6Pase-α; the resulting glucose is then transported back into the cytoplasm and released into the circulation. In the GSD-Ia liver, which lacks a functional G6Pase-α, ER-localized G6P cannot be converted to glucose, leading to hypoglycemia following a short fast.1 The GLUT2 transporter, which is responsible for the transport of glucose in and out of the cell, is shown embedded in the plasma membrane. The G6PT, which is responsible for the transport of G6P into the ER, and G6Pase-α, which is responsible for hydrolyzing G6P to glucose and phosphate, are shown embedded in the ER membrane.

Absence of Immune Response Against Human G6Pase-α.

To determine whether a humoral response directed against human G6Pase-α is generated in the infused mice, we performed Western blot analysis using the sera obtained from the 70-90-week-old control and AAV-GPE-treated G6pc−/− mice (Fig. 6E). A monoclonal antibody against human G6Pase-α that also recognizes murine G6Pase-α8 was used as a positive control (lane 1). No antibodies directed against G6Pase-α were detected in any of the infused G6pc−/− (lanes 2, 4, 6, 8, 10, and 12) or WT control (lanes 3, 5, 7, 9, 11, and 13) mice that lived to age 70-90 weeks.

Discussion

Previous short-term studies have shown that AAV-GPE completely normalizes metabolic abnormalities associated with hepatic G6Pase-α deficiency in murine GSD-Ia.8 Although these studies have shown promise, they do not address the issue of whether AAV-GPE–mediated gene transfer can prevent the development of HCA, a severe long-term complication of GSD-Ia1 and a potential complication of AAV-mediated gene therapy.11 In this study, we found that AAV-GPE–treated G6pc−/− mice expressed low levels (3%-9%, AAV-L), medium levels (22%-63%, AAV-M), and high levels (81%-128%, AAV-H) of normal hepatic G6Pase-α activity, grew normally for 70-90 weeks, displayed no hepatic abnormalities, exhibited normalized blood metabolite and glucose tolerance profiles, and had no detectable anti–G6Pase-α antibodies. The treated G6pc−/− mice exhibited no hepatic steatosis and had normal levels of hepatic triglyceride. In contrast to GSD-Ia patients1 and untreated G6pc−/− mice (this study), which cannot tolerate a short fast, the treated G6pc−/− mice sustained 24 hours of fast. Supporting this finding, significant levels of endogenous glucose were produced in the livers of AAV-GPE–treated G6pc−/− mice after a 24-hour fast, enabling them to maintain interprandial glucose homeostasis. Whereas 100% of the conditional liver-specific G6pc-null mice developed HCA at 78 weeks after gene deletion,13 none of the treated G6pc−/− mice that lived to age 70-90 weeks developed HCA. Our data demonstrate, for the first time, that AAV-GPE–mediated gene transfer not only corrects hepatic G6Pase-α deficiency on a long-term basis, but is not associated with a chronic risk of HCA.

During a fast, endogenous glucose was primarily produced in the liver via hydrolysis of G6P in the terminal step of gluconeogenesis and glycogenolysis (Fig. 7). This process is catalyzed by two ER transmembrane proteins, G6PT and G6Pase-α.1 G6PT transports G6P from the cytoplasm into the ER lumen, and G6Pase-α, with its active site inside the lumen,1 hydrolyzes intraluminal G6P to glucose (Fig. 7). These two activities are codependent, and together they maintain interprandial glucose homeostasis. When untreated G6pc−/− mice are fasted, hepatic G6P is transported by G6PT into the ER lumen, albeit at a markedly slower rate than in WT mice,12 due to the lack of G6Pase-α coupling. As a result, G6P slowly accumulates in the ER lumen. Because of G6Pase-α deficiency, no glucose is released back to the circulation, leading to fasting hypoglycemia, the hallmark of GSD-Ia1 (Fig. 7). In contrast, AAV-GPE–treated G6pc−/− mice, exhibiting 3%-128% of normal hepatic G6Pase-α activity, sustained 24 hours of fast. It is intriguing that after a 24-hour fast, the AAV-L mice expressing as little as 3% of WT hepatic G6Pase-α activity were capable of hepatic glucose production averaging 61% of WT levels. The expression of several key enzymes in gluconeogenesis (PEPCK-C, FEBase-1), glycogenolysis (PMGase), glycolysis (PFK-1), and the pentose phosphate pathway (G6PDH) (Fig. 7) were not significantly altered, but hepatic G6PT mRNA expression in AAV-GPE–treated G6pc−/− mice was 2.2-fold higher than the WT level. A similar increase in hepatic G6PT mRNA expression was also observed in untreated G6pc−/− mice (data not shown). The rate-limiting step in hepatic microsomal glucose production is G6P uptake,24 which is mediated by G6PT, and it is notable that in the presence of a reduced G6Pase-α activity there appears to be a feedback mechanism that partially compensates by increasing G6PT mRNA expression. The net result is an increase in the rate-limiting microsomal G6P uptake activity that enables AAV-L mice to maintain glucose homeostasis during prolonged fasts. This extends our understanding of the nature of functional codependence of the two components of the G6PT/G6Pase-α complex that maintains interprandial blood glucose homeostasis. In addition to a functional coupling,12 a transcriptional component may also be involved.

In AAV-GPE–treated G6pc−/− mice, G6Pase-α was distributed unevenly through the liver, with many hepatocytes exhibiting minimal activity, but foci containing markedly higher levels of activity. Although it is not clear whether the up-regulation of G6PT expression and activity is limited to specific cells, the similar increase in G6PT mRNA expression seen in untreated G6pc−/− mice suggests that it is characteristic of all hepatic cells with reduced G6Pase-α activity. The ability to compensate by feedback is a very attractive feature of this therapy, as it provides for a greater biological tolerance for the efficiency of gene therapy. Clearly, there are additional levels of control over hepatic blood glucose than recognized previously.

GSD-Ia patients under dietary therapy exhibit moderate glucose intolerance, delayed insulin response, and increased lipogenesis.26 This is most likely related to the manner of glucose intake by frequent meals of low glycemic index carbohydrate diets and/or nocturnal intragastric feeding which, by consistently elevating blood glucose levels, may lead to glucose intolerance and reduced peripheral insulin sensitivity. Insulin signaling regulates hepatic glucose and lipid metabolism that is essential for the maintenance of energy homeostasis.18 Insulin signaling associated with carbohydrate flux promotes de novo lipogenesis in the liver, but also turns off β-oxidation of fatty acids. Over the long term, this may contribute to the development of hepatic steatosis18, 27 seen in GSD-Ia patients. Consistent with this, mice with a liver-specific knockout of G6Pase-α activity are reported to develop HCAs and exhibit marked hepatic steatosis.13 In contrast, the AAV-GPE–treated G6pc−/− mice in our study did not manifest hepatic steatosis and had normal hepatic triglyceride. Both WT and the treated G6pc−/− mice exhibited normal fasting blood insulin levels, although the treated G6pc−/− mice with lower fasting blood insulin levels indicate a better metabolic control. In AAV-GPE–treated G6pc−/− mice, hepatic free glucose levels, an indicator of endogenous glucose production, were consistently lower than those in WT mice, leading to reduced fasting blood insulin in the treated mice. Supporting this, fasting blood insulin levels in the treated G6pc−/− and control mice correlated with their body weights, and the treated G6pc−/− mice were considerably leaner. The low fasting blood insulin levels and lack of hepatic steatosis may explain why AAV-GPE–treated G6pc−/− mice do not develop HCA.

In conclusion, we have demonstrated that AAV-GPE–mediated gene therapy in G6pc−/− mice that restores 3%-128% of normal hepatic G6Pase-α activity corrects hepatic G6Pase-α deficiency and is not associated with HCA formation. Our results suggest that G6Pase-α gene transfer may offer a therapeutic approach to the management of human GSD-Ia.

Ancillary