Phenotypic Correction of Diabetic Mice by Adenovirus-Mediated Glucokinase Expression

The 1.4-kb full-length cDNA encoding human liver GK (21) was polymerase chain reaction (PCR)-amplified from a human liver cDNA library (Clontech, Palo Alto, CA). PCR was carried out using 1 ng of the library and primers containing EcoRI and SalI sites (upstream primer 5′-GAATTCATGGCGATGGATGTCACAAGG-3′; downstream primer 5′-GTCG ACTCACTGGCCCAGCATACAGG-3′), using Clontech’s Advantage-GC cDNA PCR kit according to the manufacturer’s instructions. After the reaction was heated for 1 min at 94°C, PCR was performed for 30 s at 94°C followed by 3 min at 68°C for 35 cycles with a final extension for 3 min at 68°C. The resulting PCR product was purified from an agarose gel using the QIAEX II kit (Qiagen, Valencia, CA) and then ligated into the TA-cloning vector, pCR2.1 (Invitrogen, Carlsbad, CA). The cDNA insert sequence was verified using ABI prism dye terminator cycle sequencing on a Model 377 DNA Sequencer (PE Biosystems, Foster City, CA). This plasmid was digested with SalI, Klenow treated, and then digested with EcoRI. A double-stranded EcoRI adapter (5′-CTAGCCACCCACCCC-3′ and 5′-AATTGGGGTGGGTGG-3′) containing an overhang complementary to the SpeI site of the adenovirus shuttle vector pAvS6a (22) was ligated to the EcoRI site of the GK cDNA fragment. The resulting fragment was ligated into the pAvS6a digested with SpeI and EcoRV to generate pAvhGK. A lox site (5′-ATAACTTCGTATAATGTATGCTATACGAAGTTAT-3′) was added downstream of the poly(A) signal to generate pAvhGKlx. The GK expression cassette included a constitutive RSV promoter, 198-bp fragment containing the adenovirus serotype 5 major late promoter tripartite leader sequence and an SV40 early polyadenylation signal.

The recombinant adenovirus encoding human GK (Av3hGK) was constructed by a rapid vector generation protocol using Cre recombinase-mediated recombination (23) of two plasmids, one containing the right-hand portion of the adenoviral vector genome and a lox site and the other plasmid, pAvhGKlx, containing the left end of the viral genome, the GK expression cassette, and a lox site. Both plasmids along with a Cre-encoding plasmid were cotransfected into AE1–2a cells, which are A549 cells stably transfected with the adenovirus E1/E2a regions under dexamethasone-inducible promoters (22). DNA from the recombinant vector was isolated and analyzed by restriction digestion. Large-scale vector preparations were made, and the vector concentrations were determined by spectrophotometric analysis (24). Titers are stated as particles per milliliter. The control null vector (Av3Null) was identical to the GK vector except that it lacked a transgene, but it retained the RSV promoter and SV40 poly(A) signal.

Three- to 4-week-old C57BL/6J male mice were purchased from Jackson Laboratories (Bar Harbor, ME). All mice were housed in a pathogen-free barrier facility and were maintained on a 12-h light/dark cycle. These mice were maintained on either a high-fat diet (HF; 58.0% fat, 25.6% carbohydrates, and 16.4% protein calories; diet # D12309R) or a control diet (LF; 10.5% fat, 73.1% carbohydrates, 16.4% protein calories; diet # D12310; Research Diets, New Brunswick, NJ) for at least 4 months (20), after which diabetic mice were selected on the basis of elevated fasting blood glucose levels and impaired glucose tolerance. Diabetic mice were randomly divided into three treatment groups (Hank’s balanced salt solution [HBSS], Av3Null, or Av3hGK) such that each group had the same average fasting blood glucose levels. Adenoviral vector administrations via tail-vein injections and retro-orbital phlebotomy were performed as described (25). Previous studies demonstrated that intravenous administration of adenoviral vectors to mice results in efficient liver transduction with lung, spleen, and skeletal muscle being much less efficiently transduced (14). Normal LF mice received 6 × 10¹⁰ particles/animal, and the larger HF mice received 12 × 10¹⁰ particles/animal. All vector dilutions were done in HBSS, and the injection volume was 250 μl. These doses resulted in similar liver transduction efficiencies in the LF and HF mice (Fig. 1). The diet treatment was continued after the vector administration until study termination.

OGTTs were performed 2 weeks before and 2, 7, and 12 weeks after the vector treatment. The OGTT protocol was as follows. The mice were fasted for 14–16 h. The fasting body weights were recorded, and blood samples were collected in Fl/EDTA tubes (Sarstedt, Newton, NC). A glucose bolus (1 g/kg body wt as a 10% wt/vol solution) was administered by oral gavage. Blood samples were then collected at 30 and 120 min after oral glucose administration for glucose and insulin measurements. At weeks 3, 7, and 13 after vector treatment, animals were killed by cervical dislocation, the livers were harvested, and the liver weights were recorded. Tissue sections were either fixed in formalin for hematoxylin and eosin (H&E) stain or snap-frozen in liquid nitrogen. All animal procedures were conducted in accordance with principles and guidelines established by the Institutional Animal Care and Use Committee in accordance with the Animal Welfare Act.

For food intake and body weight studies, the mice were housed with bedding in individual plastic shoebox cages with metal wire basket tops. HF food was placed on the basket top (as normal), and the basket plus food was weighed as one (food-IN). Every 2–3 days, the basket plus food was reweighed (food-OUT), and food intake was calculated (INTAKE = [IN-OUT]/[# days]). The cage bedding was carefully scanned for any food pellets that may have fallen through the basket top; these were then added to the basket and included in the weight. When there was a large amount of food pellets in the bedding that precluded an accurate measurement, the food intake for that mouse was not calculated on that day. Additional food was added and included in the food-IN value when necessary. Water was provided ad libitum. Mice were individually weighed at the same time every 2–3 days.

Blood glucose concentrations were measured using a hand held glucometer (Bayer, Tarrytown, NY). Plasma insulin was measured using an enzyme-linked immunosorbent assay kit from Crystal Chem (St. Louis, MO) with mouse insulin as a standard. Plasma samples were sent to an outside laboratory (AniLytics, Gaithersburg, MD) for determination of glucose, TG, FFA, lactate, and alanine transaminase (ALT).

Liver GK activity was determined spectrophotometrically from the liver extracts. Briefly, 200–300 mg of liver was added to 500 μl of RIPA buffer (50 mmol/l Tris-HCl [pH 7.5], 150 mmol/l NaCl, 2 mmol/l EDTA, 1 mmol/l EGTA, 0.1% Triton X-100, 10% glycerol) containing 10 μg/ml leupeptin, 10 μg/ml aprotinin, 25 μg/ml pefabloc, and 1 mmol/l phenylmethylsulfonyl fluoride. The tissue was homogenized (Kontes Duall tissue homogenizer) on ice for 7–10 s and then sonicated on ice for 5 s (twice). The sample was incubated for 10 min on ice and then centrifuged at 15,000g at 4°C for 5–10 min. The supernatant was removed and stored at −80°C until use or diluted 1:10 in buffer (100 mmol/l Tris-HCl [pH 7.4], 100 mmol/l KCl, 6 mmol/l MgCl₂) and assayed for GK activity using a method essentially as described by Hariharan et al (12), except that the assay buffer contained 100 mmol/l Tris-HCl [pH 7.4], 100 mmol/l KCl, 6 mmol/l MgCl₂, 1 mmol/l DTT, 5 mmol/l ATP, 1 mmol/l thioNAD, 30 units/ml glucose-6-phosphate dehydrogenase, and 0.5 or 100 mmol/l glucose. GK activity was estimated as the difference in activity when samples were assayed at 100 mmol/l (GK plus hexokinase activity) and 0.5 mmol/l glucose (hexokinase activity).

Hepatic lipid analysis was performed after homogenizing liver tissue biopsies weighing between 20 and 200 mg in HPLC grade acetone at a ratio of 1:20 (wt:vol). Tissue homogenates were shaken overnight to allow complete lipid extraction. Aliquots of the lipid extract were analyzed for TG using an Infinity Triglycerides Reagent Kit (Sigma, St. Louis, MO).

Liver samples were homogenized in 0.03 N HCl (to a final concentration of 0.5 g/ml). The homogenate (100 μl) was mixed with 400 μl of 1.25 N HCl and heated for 1 h at 100°C. Samples were centrifuged at 14,000 rpm, and 10 μl of supernatant was mixed with 1 ml of glucose oxidase reagent (Sigma). After a 10-min incubation at 37°C, the absorbance was read at 505 nm. A standard curve using glycogen type III obtained from rabbit liver (Sigma) was also simultaneously analyzed to determine the final liver glycogen concentrations.

The liver extract for the Western blot analysis was prepared as described above in gk activity measurements. Determination of the protein concentrations and SDS-PAGE were performed using standard protocols. The protein was transferred to a nitrocellulose membrane and blocked overnight with Tris-buffered saline (TBS)/bovine serum albumin (BSA) (10 mmol/l Tris-HCl [pH 7.4], 150 mmol/l NaCl, 0.2% Tween-20, 3% BSA). The membrane was incubated with the primary antibody (rabbit anti-GK Santa Cruz H-88; 1:50 dilution; Santa Cruz Biotechnology, Santa Cruz, CA) for 2 h with continuous rocking at room temperature. The membrane was washed two times for 10 min and blocked for 10 min with TBS/BSA followed by a 2-h incubation with the secondary anti-rabbit horseradish peroxidase–conjugated antibody (Promega, Madison, WI; 1:25,000). After three 10-min washes, the membrane was treated with the enhanced chemiluminescence reagent (Amersham Pharmacia Biotech, Piscataway, NJ) and immediately exposed to film.

DNA was isolated from mouse livers using the QIAmp Tissue Kit (Qiagen, Chatsworth, CA). Ten micrograms of each DNA sample was digested with SphI and subjected to Southern blot analysis (25). The probe, prepared by random oligonucleotide priming, was a purified 1,475-bp SphI fragment isolated from a plasmid containing adenoviral vector sequences from bp 7,790–9,464. The copy number control standard was prepared by adding 600 pg of viral DNA, equivalent to 10 vector copies per cell, to 10 μg of control mouse liver genomic DNA.

Results are reported as mean ± SE. The comparison of different groups was carried out using unpaired Student’s t test. Differences were considered statistically significant at P < 0.05.

To examine the effect of GK overexpression in a murine model of type 2 diabetes, we maintained normal male C57BL/6J mice on an HF diet for at least 4 months (20). The HF mice displayed higher body weights (42 ± 0.8 g, elevated fasting blood glucose (157 ± 4.3 mg/dl), and plasma insulin levels (1,826 ± 214 pg/ml) as compared with the age-matched control mice maintained on a normal, LF diet. LF mice had lower body weight (28.5 ± 0.8 g) and displayed normal fasting blood glucose (98.8 ± 6.9 mg/dl) and insulin (726 ± 149 pg/ml) levels. The HF mice also displayed an impaired glucose tolerance.

The adenoviral vector Av3hGK contains an RSV promoter, a 198-bp adenovirus serotype 5 major late promoter tripartite leader, the human hepatic GK cDNA, and an SV40 early polyadenylation signal. The vector backbone was derived from adenovirus serotype 5 (Ad5) and was devoid of E1/E2a and E3 regions (22). LF mice received 6 × 10¹⁰ particles/animal, and the larger HF mice received twice that dose (1.2 × 10¹¹ particles/animal) of Av3hGK, Av3Null, or HBSS. These vector doses resulted in similar liver transduction efficiencies (2–4 vector copies per cell) as determined by Southern blot analysis (Fig. 1).

To study the effect of overexpression of hepatic GK on the fasting blood glucose and insulin levels in the HF and LF mice, we collected blood samples from overnight-fasted animals that were treated with Av3hGK, Av3Null, or HBSS. Blood glucose analyses of the samples revealed that the Av3hGK-treated diabetic mice had a 24.5% reduction in blood glucose levels as compared with control groups as early as week 1 after vector treatment (Table 1). The Av3hGK-treated diabetic mice also displayed a 64.4% reduction in fasting plasma insulin levels as compared with control groups at week 1 posttreatment (Table 2). The vector treatment did not have any significant effect on either the fasting blood glucose or plasma insulin levels in the age-matched LF control mice during the course of the experiment (Tables 1 and 2). No changes in basal glycemia and insulin levels after administration of a similar dose of Av3hGK to LF mice was observed in two additional studies (data not shown).

The fasting blood glucose levels in the HF diabetic mice that were treated with Av3hGK was lowest at week 2 posttreatment and remained low until the end of the experiment at week 13 posttreatment (Table 1). Thus, at 13 weeks, the Av3hGK group continued to show a 28.8% reduction in fasting blood glucose as compared with the vehicle-treated control group and a 5% reduction compared with the Av3Null group (Table 1). The Av3hGK group also displayed a 28.9% reduction in plasma insulin levels as compared with the HBSS-treated control group and an 18.6% reduction compared with the Av3Null group at week 12 (Table 2). Therefore, Av3hGK treatment in the HF diabetic mice resulted in normalization of fasting hyperglycemia and hyperinsulinemia, and the effect lasted for several weeks.

To study the effect of hepatic overexpression of GK on glucose disposal, we subjected the HF and LF mice to an OGTT 2 weeks after administration of Av3hGK, Av3Null, or HBSS. The fasting blood glucose levels in the Av3hGK-treated diabetic mice were 28.9% lower (P < 0.00001) compared with control groups (Fig. 2A). More importantly, the glucose levels at the 30-min time point in the Av3hGK group were 42% lower (P < 9 × 10⁻⁶) compared with the control groups (39% reduction in the area under the curve; P < 0.01), suggesting a faster glucose clearance rate. Notably, the glucose tolerance of the Av3hGK-treated mice was similar to that of the normal, LF-treated mice (Fig. 2A). Thus, expression of GK in the diabetic mice resulted in complete normalization of glucose tolerance.

Measurement of plasma insulin levels revealed that Av3hGK treatment resulted in a significant reduction in fasting insulin levels in the diabetic mice (66%) compared with both of the control groups (Fig. 2B). The insulin levels in the HF Av3hGK treatment group remained in the normal range during the course of the OGTT with a 70% reduction at time 30 min and a 72% reduction at time 120 min after oral glucose administration as compared with the control groups (Fig. 2B). These data suggest that the normalization of glucose tolerance may be the result of increased hepatic glucose uptake mediated by increased hepatic GK activity. This effect of a significantly increased rate of glucose clearance during the glucose tolerance test in the Av3hGK-treated group lasted for at least 7 weeks after vector administration, with some effect still evident at 12 weeks after vector administration (data not shown). The Av3hGK treatment did not have any significant effect on the glucose tolerance of the age-matched LF mice (Fig. 2A).

To quantify total GK protein and activity in the experimental mice, three mice were killed from each treatment group at 3 and 13 weeks after vector (or HBSS) administration. Western blot analysis revealed that Av3hGK-treated mice exhibited an increase in the GK protein levels when compared with the Av3Null- or HBSS-treated mice at both 3 and 13 weeks (Fig. 3). Similarly, the GK activity levels in the livers of Av3hGK-treated HF mice showed a 119% increase as compared with the levels seen in HBSS-treated HF mice (Table 3). Similar increases in the hepatic GK activity after Av3hGK treatment were observed in three previous analogous studies with cohort size of at least 10. It is interesting that despite a 72% increase in the hepatic GK activity (Table 3), Av3hGK-treated age-matched LF mice did not display any significant difference in plasma glucose or insulin levels.

To study the effect of overexpression of GK on various blood parameters, we collected blood samples from overnight-fasted mice at week 3 posttreatment. Whereas moderate GK overexpression in HF mice (Table 3) resulted in correction of the diabetic phenotype (Fig. 2), no significant effects on circulating FFA or TG levels were observed in these mice (Table 4). However, an increase in the plasma ALT levels in the Av3hGK-treated groups (both HF and LF) was observed. This increase in the ALT levels can be attributed to the adenovirus vector treatments (14) as the Av3Null treatment also resulted in a similar increase in ALT levels. It is interesting that no significant increase in the plasma lactate level was observed in the Av3hGK-treated HF or LF animals (Table 4).

In a separate study, hepatic glycogen and hepatic fat content was assessed in treated diabetic animals. In this study, HF mice were treated with either 1.2 × 10¹¹ particles of Av3Null (n = 12) or a mixture of 6 × 10¹⁰ particles each of Av3hGK (n = 10) and Av3Null (n = 10), keeping the total input virus dose of the groups constant. Livers were harvested 3 weeks after vector treatment. GK activity in the livers of Av3hGK-treated animals was significantly increased (Table 5). It is interesting that these levels were approximately half those measured in animals that received a two-fold higher GK vector dose (Table 3). Furthermore, blood glucose and insulin levels were reduced, and glucose tolerance was normalized in these animals (data not shown), similar to animals that received the two-fold higher Av3hGK vector dose (Tables 1 and 2 and Fig. 2). Av3hGK treatment resulted in a 70% increase in liver weight compared with the HBSS control group, whereas Av3Null vector treatment resulted in a 40% liver weight increase (Table 5). Calculation of liver weight as percentage of body weight revealed that Av3hGK-treated mice displayed a 30% increase compared with the Av3Null-treated group (Table 5). Possible explanations for this observed increase in liver weight include an increase in glycogen and/or fat content. Analyses of hepatic glycogen revealed a dramatic, 175% increase in GK-treated animals, whereas Null treatment resulted in a 101% increase compared with vehicle-treated animals (Figs. 4A and Table 5). Quantitation of hepatic TG content revealed no significant differences among the GK, Null, or HBSS-treated animals (Fig. 4B and Table 5).

Liver sections obtained from treated animals were analyzed by H&E staining (Fig. 5). An increased number of vacuoles within hepatocytes were observed in HBSS alone–treated HF mice compared with the normal LF mice (Figs. 5A and D). Av3hGK treatment resulted in localized pockets of large vacuoles in the hepatocytes (Figs. 5C and F). Av3Null treatment also resulted in increased vacuolization (Figs. 5B and E); however, the vacuoles were smaller and more scattered across the liver sections. It remains unclear what these vacuoles contained; however, it is possible that they contained glycogen. Analysis of the H&E sections by a pathologist (J. Markovits, Novartis, Summit, NJ) revealed an increase in liver glycogen content of both the Av3hGK- and Av3Null-treated groups. Five Av3hGK-treated animals (n = 10) showed marked glycogen accumulation, whereas four displayed moderate accumulation, and one animal showed slight accumulation. The Av3Null treatment group (n = 10) showed two animals with marked, five with moderate, and three with slight glycogen accumulation. The HBSS-treated group (n = 10) displayed one animal with moderate, three animals with slight, and six with minimal glycogen accumulation. It is interesting that there was no detectable difference in overall liver lipidosis among the treatment groups.

To assess the effect of Av3hGK treatment on food intake and body weight, HF mice were treated with either 1.2 × 10¹¹ particles of Av3Null (n = 12) or a mixture of 6 × 10¹⁰ particles each of Av3hGK and Av3Null, keeping the total input virus dose of the groups the same (n = 10). HBSS-treated animals were not assessed in this study. Food intake and body weight were measured at the indicated times (Fig. 6). Mice that were treated with Av3hGK displayed a significant reduction in food intake at days 13, 18, and 21. Consistent with this finding, the Av3hGK-treated animals exhibited a steady drop in body weight to a level, at 3 weeks, that was significantly less than both their pretreatment weights (9.98% reduction; P < 0.005) and to Av3Null-treated mice (9.64% reduction; P < 0.05).

Hepatic glucose uptake and utilization is defective in type 2 diabetes, and it has been suggested that increasing the hepatic GK activity may lead to improved glucose utilization by the liver and thereby improve the whole-body glucose homeostasis. This work represents the first evaluation of GK overexpression via gene therapy in a diabetic animal model. We demonstrated that the treatment of diabetic mice with an adenoviral vector encoding human hepatic GK resulted in normalization of fasting blood glucose and insulin levels and normalization of glucose tolerance compared with an age-matched nondiabetic control group. Furthermore, no increase in plasma or hepatic TG or in plasma FFA was observed.

Treatment of diabetic mice with Av3hGK resulted in a 2.2-fold increase in liver GK activity levels. An increase in enzymatic activity of the rate-limiting step in hepatic glucose metabolism should allow for increased clearing by the liver of glucose from the blood. Consistent with this hypothesis, this small increase in the hepatic GK activity was sufficient to normalize the fasting hyperglycemia of the diabetic mice. The lowering of the fasting blood glucose levels was observed as early as 1 week after the intravenous administration of the GK vector and lasted for at least 13 weeks. Thus, relatively long-lasting expression was observed after a single intravenous injection of Av3hGK. Attenuation of E2a gene expression has been reported to prolong transgene expression in some animal models (26,27,28). In fact, the expression of human factor VIII using similar third-generation adenoviral vectors in a hemophiliac mouse model (also C57BL/6 background) was shown to last for up to 1 year (29).

Normalization of fasting hyperglycemia in the Av3hGK-treated diabetic mice was associated with a significant reduction in fasting plasma insulin levels, suggesting improved insulin sensitivity. These data are consistent with the work of Postic et al. (8) in which the absence of hepatic GK activity in the liver-specific GK knockout mice resulted in hyperinsulinemia, suggesting that the lack of hepatic GK may cause mild insulin resistance.

The effect of the Av3hGK vector treatment was more pronounced when the diabetic mice were challenged with an oral glucose load. The normalization of glucose tolerance in the Av3hGK-treated diabetic mice can be attributed to increased hepatic GK activity achieved by the Av3hGK vector. Indeed, the Av3hGK-treated mice displayed lowered insulin levels during the OGTT, which paralleled that of the normal LF control mice. It is interesting to speculate that the flat insulin curve (Fig. 2B) displayed by HF GK-treated mice was a result of the flattened glucose curve (Fig. 2A), suggesting that the rapid removal of glucose from the bloodstream may have prevented the release of insulin by the pancreas.

It was reported previously that low-level, adenovirus-mediated GK expression did not affect the fasting blood glucose levels in normal rats (17), suggesting that a compensatory mechanism prevented the animals from becoming hypoglycemic. However, in the normal rats, expression of GK at high levels resulted in mild hypoglycemia with increased plasma FFA and TG levels (17). Thus, a very high dose of the GK vector was needed to see any effect on the circulating levels of metabolites in normal rats, which may have perturbed the glucose metabolism in the animals and resulted in hyperlipidemia. Indeed, in our studies using normal, nondiabetic animals, we obtained similar results. Treatment of normal mice with Av3hGK resulted in no effect on fasting blood glucose levels (Table 1), whereas treatment of normal mice with a five-fold higher dose of Av3hGK, resulting in a 40-fold increase in hepatic GK expression, resulted in hyperlipidemia (U.J.D., manuscript in preparation). In contrast, in the current study, a moderate 1.5- to 2-fold increase in hepatic GK activity in diabetic animals was enough to normalize glucose homeostasis without affecting plasma lipid homeostasis. Consistent with these results, Shiota et al. (30) recently reported that transgenic mice expressing one extra copy of the GK gene locus exhibited a dramatic reduction in the development of hyperglycemia and hyperinsulinemia induced by an HF diet compared with nontransgenic mice. It is interesting that adenoviral overexpression of GK in HF mice also led to a reduced food intake and body weight (Fig. 6) and a significant reduction in interscapular brown adipose, retroperitoneal, and perirenal fat pad mass compared with the Av3Null-treated group (data not shown). Whereas some of the effects of GK vector treatment could be attributed to this decreased food intake, the observed increase in hepatic glycogen content could not be caused by reduced food intake. Taken together, these data suggest that a moderate increase in GK activity can affect whole-body carbohydrate and lipid metabolism and lead to changes in food intake. Thus, the hepatic GK level seems to be crucial in whole-body energy homeostasis.

Av3hGK-treated mice showed mild hepatomegaly associated with increased hepatic glycogen storage, which was probably caused by improved glucose uptake by the liver. It is unclear why Av3Null vector treatment also caused an increase in glycogen accumulation and may be related to the observed mild vector-induced toxicity (Tables 2 and 5). It is interesting that quantitative and empirical analyses of hepatic TG content did not show a difference among the treatment groups. However, H&E analysis of liver sections showed pockets of vacuolization that seemed to be increased in the HF Av3hGK-treated group. It remains unclear what was contained within these vacuoles, but the observed increase in hepatic glycogen suggests that the vacuoles may have contained glycogen. Alternatively, adenoviral vector treatment of the mice was associated with mild liver toxicity. This toxicity may also have contributed to the observed liver vacuolization. The toxicological issues associated with adenovirus-mediated gene expression can be addressed by the use of gutless adenoviral vectors. Gutless vectors, which are devoid of all viral genes, have been shown to be significantly less toxic and immunogenic than early generation vectors (31). It will be interesting to determine whether the use of a gutless adenoviral vector encoding GK will circumvent the observed vector-induced toxicity and perhaps the liver vacuolization issues while mediating correction of the diabetic phenotype.

We thank Dr. Theodore Smith and Jimmy Zhao for help with the animal experiments, Christine Mech for the histological analyses, Dr. Judit Markovits for pathological examination of the liver sections, Kathy Ramos for the hepatic glycogen and fat analyses, and Dr. Börk Balkan for helpful discussions. Special thanks to Dr. Gene Liau for critical review of the manuscript.