Skip to main content
NCBI home page
Molecular Therapy logoLink to Molecular Therapy
. 2016 Mar 8;24(4):697–706. doi: 10.1038/mt.2016.35

In Vivo Zinc Finger Nuclease-mediated Targeted Integration of a Glucose-6-phosphatase Transgene Promotes Survival in Mice With Glycogen Storage Disease Type IA

Dustin J Landau 1,2, Elizabeth Drake Brooks 1,3, Pablo Perez-Pinera 4,5, Hiruni Amarasekara 1,6, Adam Mefferd 2, Songtao Li 1, Andrew Bird 1, Charles A Gersbach 4, Dwight D Koeberl 1,2,*
PMCID: PMC4886939  PMID: 26865405

Abstract

Glycogen storage disease type Ia (GSD Ia) is caused by glucose-6-phosphatase (G6Pase) deficiency in association with severe, life-threatening hypoglycemia that necessitates lifelong dietary therapy. Here we show that use of a zinc-finger nuclease (ZFN) targeted to the ROSA26 safe harbor locus and a ROSA26-targeting vector containing a G6PC donor transgene, both delivered with adeno-associated virus (AAV) vectors, markedly improved survival of G6Pase knockout (G6Pase-KO) mice compared with mice receiving the donor vector alone (P < 0.04). Furthermore, transgene integration has been confirmed by sequencing in the majority of the mice treated with both vectors. Targeted alleles were 4.6-fold more common in livers of mice with GSD Ia, as compared with normal littermates, at 8 months following vector administration (P < 0.02). This suggests a selective advantage for vector-transduced hepatocytes following ZFN-mediated integration of the G6Pase vector. A short-term experiment also showed that 3-month-old mice receiving the ZFN had significantly-improved biochemical correction, in comparison with mice that received the donor vector alone. These data suggest that the use of ZFNs to drive integration of G6Pase at a safe harbor locus might improve vector persistence and efficacy, and lower mortality in GSD Ia.

Introduction

Glycogen storage disease type Ia (GSD Ia), also known as von Gierke's disease, is caused by mutations in the G6PC gene encoding glucose-6-phosphatase (G6Pase). Loss of functional G6Pase prevents conversion of glucose-6-phosphate to glucose, leading to myriad symptoms, chief among them chronic and potentially lethal hypoglycemia. Current treatment aims at preventing hypoglycemia through a strict dietary regimen. However, despite treatment, these patients still experience other consequences of glycogen accumulation such as hepatic adenomas and renal failure.1

AAV vectors have been widely used for preclinical gene therapy for GSD Ia and other diseases requiring transgene expression in the liver.2,3,4,5,6,7,8 We and others found that AAV vector administration to young animals accomplished a high level of liver transduction, however this is followed by declining numbers of vector genomes over the ensuing months due to hepatocyte cell division that is especially high in young GSD Ia mice that must replenish lost hepatocytes due to high turnover rates.9,10,11,12 In particular, Chou and colleagues highlighted the transience of AAV vector-mediated expression by measuring a 12-fold decline in G6Pase expression in mice with GSD Ia from age 2 to 6 weeks.12 Over 90% of G6Pase introduced by an AAV2/8 vector was lost between 6 and 18 months following vector administration in a later study by the same group.13 We demonstrated that the GSD Ia liver features a high rate of apoptosis associated with the loss of AAV vector genomes.14 These studies show that although AAV vectors confer significantly longer expression than other common episomal gene therapy vectors, including adenovirus or plasmid DNA vectors, expression from an AAV vector in the liver does decrease over time.15,16 As a result, repeated administration of gene therapy vectors may be necessary with these strategies. Although adaptive immune response to the viral capsid proteins prevents readministration of the same serotype, repeated dosing is possible due to the previous description of many distinct AAV capsid serotypes useful for gene therapy,17 which can extend therapeutic efficacy. However, this approach is impractical for treatment of human patients since expression of the transgene for multiple decades is essential. One solution for GSD Ia gene therapy is therefore an approach that integrates the G6Pase gene into the genome, such that dividing hepatocytes will retain G6Pase expression. This would be able to provide permanent treatment after a single dose of gene therapy vector.

Modern gene-editing tools permit integration of DNA sequences at targeted loci through homologous recombination, which is facilitated by introduction of a double-strand break in genomic DNA.18 This strategy has been implemented to correct hemophilia B in mouse models.19,20 The present study was designed to expand the interval of clinical benefit and to enhance the response to gene therapies for GSD Ia by integrating G6Pase at the ROSA26 locus by codelivering ZFNs21 and a targeting vector using AAV.

In our study, we observed enhanced survival past 32 weeks in mice that received the ZFN, compared to less than 22 weeks without ZFN, as well as target site-specific integration events corresponding to ZFN administration. These results confirm that this approach works as hypothesized to increase long-term efficacy from an AAV vector in GSD Ia. This proof-of-concept work integrating a G6Pase transgene into a safe-harbor locus is a significant step toward single-dose permanent treatment for an otherwise progressive life-threatening disease.

Results

In vivo ZFN activity following AAV transduction in mice

The vector AAV-ZFN contained an expression cassette consisting of the thyroid hormone-binding globulin liver-specific promoter (LSP)22 driving high-level expression of the ROSA26-specific ZFN21 in the liver. The vector AAV-RoG6P contained the cDNA for human G6Pase under control of the minimal human G6Pase promoter4 and a bovine growth hormone polyA signal, flanked on each side by ~850 bp sequences with homology to the ROSA26 target site. These vectors were initially cross-packaged with the AAV8 capsid, and later with the AAV9 capsid (Table 1) to accomplish high-level liver transduction.

Table 1. Mouse treatment groups.

graphic file with name mt201635t1.jpg

Open in a new tab

To assess the activity of the ROSA26 ZFN at its target site in vivo, 2-week-old G6pc-/- mice were injected with AAV2/8-ZFN and AAV2/8-RoG6P, each at a dose of 2E+13 viral particles per kilogram (Table 1). Genomic DNA from the livers of these mice at 8 months of age was used to determine ZFN activity at the target site by the Surveyor assay.23 DNA cleavage sites are typically repaired using error-prone nonhomologous end-joining (NHEJ), which creates indels at genomic sites where repair takes place. The average allele modification rate measured in knockout mice (n = 4) was ~2% whereas in wild-type mice (n = 3) it was ~0.44% (Figure 1). This estimate does not include alleles in which targeted donor vector integration occurred. There was no evidence of ROSA26 cleavage in mice that did not receive AAV-ZFN (n = 3) (Figure 1). These experiments demonstrated that the ROSA26-ZFN is active in mouse liver when delivered in vivo as an AAV2/8 vector, with slightly higher levels of allele modification in G6pc-/- mice.

Figure 1.

Figure 1

Open in a new tab

Gene editing at the ROSA26 locus occurs specifically in mice administered AAV-ZFN/8. (a) The Surveyor assay demonstrated that indels indicative of nonhomologous end-joining occurred only in mice that received AAV-ZFN/8. G6pc-/- knockout mice had significantly more ZFN target site-specific DNA repair events than wild-type mice. (b) Representative gel images illustrating the difference between knockout and wild-type mice. Surveyor nuclease-digested product bands are indicated by black arrows. “WT, UT”: untreated wild-type mouse control, representative of the control mice that did not receive AAV2/8-ZFN. “-n”: no-Surveyor nuclease control of wild-type mouse PCR product. Error bars: mean ± SD.

Therapeutic transgene longevity in a GSD Ia mouse model via AAV delivery

Since the goal of this experiment is to extend the persistence of a therapeutic transgene by integration into the mouse genome at a target locus, we confirmed the presence of the G6Pase transgene at the targeted ROSA26 locus in the above-mentioned 8-month-old mice injected at 2 weeks of age. Two rounds of polymerase chain reaction (PCR) were required to amplify junctions between the ROSA26 locus and human G6Pase transgene from the AAV2/8-RoG6P vector (Figure 2). We confirmed the identity of the anticipated junction by subcloning the PCR products and Sanger sequencing. The predicted junction was detected in 7 of 10 mice treated with both AAV2/8-RoG6P and AAV2/8-ZFN vectors. Only one of eight non-ZFN-treated mice yielded significant PCR amplification, but the yield was too low to sequence. Assuming that this single mouse is a true nuclease-independent targeted integration event, perhaps as a result of low level ZFN-independent homologous recombination,24 there is still a significant difference between the two groups with P < 0.03 (Fisher's two-tailed test). One double-vector-treated mouse that contained a homology-directed repair junction also contained a separate NHEJ integration event consisting of the entire donor vector at the ROSA26 site (data not shown).

Figure 2.

Figure 2

Open in a new tab

Vector transgenes are integrated into the ROSA26 locus. G6pc-/- knockout and wild-type mice were treated with both AAV2/8-ZFN and AAV2/8-RoG6P or only AAV2/8-RoG6P prior to analysis at 8 months of age. (a) Illustration of the predicted integration structure following homology-directed repair between the mouse genomic ROSA26 cleavage site and the AAV2/8-RoG6P viral genome donor homology arms. Primer locations are denoted P1-P4. (b) Representative gel of PCR products from all mice following the nested round of PCR. Predicted product size from an HDR event is 1,335 bp. Ladder is 1 kb Plus (Invitrogen, Waltham, MA). White arrow indicates the 1 kb position. “PBS”: PCR reaction was run on DNA from an affected mouse injected with PBS instead of vector and collected at 2 weeks of age. “NTC”: No template control PCR reaction, run using water instead of DNA as template.

To determine whether integration had the desired effect of increasing transgene longevity, hepatic transgene levels in knockout mice were quantified by qPCR. Mouse samples were compared to a standard curve as previously described8 to calculate the number of AAV2/8-RoG6P and AAV2/8-ZFN genomes present per cell (Figure 3a). We found no significant difference in G6Pase transgene copy number between the dual-vector and single-vector groups. However, there was a significant difference of P < 0.02 between the G6Pase and ZFN copy numbers within the dual-vector group, showing 13-fold lower retention of the AAV2/8-ZFN genomes. A lower level of retention of the ZFN vector may reduce the risk of potential off-target genome modification that could result from long-term ZFN expression. Additionally, analyses of mice at 3 months of age that had been administered either one or both vectors cross-packaged as AAV9 at adjusted dosages (Table 1) did not reveal a difference in the G6Pase transgene copy number between groups (Figure 3b).

Figure 3.

Figure 3

Open in a new tab

Donor vector genomes persist better than ZFN genomes in hepatocytes. (a) AAV2/8-RoG6P and AAV2/8-ZFN genomes were quantified by qPCR of G6pc-/- mouse liver DNA at 8 months of age. The ZFN genome was found to exist at much lower levels than the RoG6P genome when both were delivered in equal amounts by AAV vectors (P < 0.014). (b) AAV2/9-RoG6P genomes were quantified by qPCR of knockout mouse liver DNA at 3 months of age. The G6Pase transgene levels did not differ between treatment groups. Error bars: mean ± SD.

Virus-mediated therapeutic effects in a GSD Ia mouse model

Untreated G6pc-/- mice are typically incapable of surviving longer than 3 weeks.9 We have previously established that treatment with a G6Pase cDNA transgene under control of the minimal human G6Pase promoter increases mouse survival to over 12 months.8 To determine whether coadministration of a ZFN cleaving a safe harbor locus and a targeted donor transgene could improve the therapeutic duration in mice, we compared the survival rates of G6pc-/- mice that were coadministered AAV2/8-ZFN and AAV2/8-RoG6P (n = 6) with those of mice that received the AAV2/8-RoG6P donor vector alone (n = 7). Dual-vector mice significantly outlived single-vector mice (P < 0.04), and the difference was particularly noticeable early in life (Figure 4a). This suggests improved therapeutic duration when coadministering an AAV-ZFN that catalyzes targeted transgene integration. Furthermore, survival analysis of only female mice demonstrated uniform survival in the coadministration group, in contrast with 67% mortality in the AAV2/8-RoG6P alone group (Figure 4b; P < 0.05). Mortality among female mice specifically was prevented by administration of AAV-ZFN. The mortality among female mice might be attributed to decreased transduction with an AAV vector, in comparison with male mice.25

Figure 4.

Figure 4

Open in a new tab

Prolonged survival of G6pc-/- mice depends on administration of AAV2/8-ZFN. (a) All mice that received AAV-ZFN/8 in addition to AAV-RoG6P/8 survived for 8 months. In contrast, only 43% of mice that received AAV-RoG6P/8 alone survived for the same duration, showing a significant difference (P < 0.04, using the log-rank test). (b) Female mice showed the same requirement for ZFN as the population as a whole, with 100% survival among mice receiving both vectors and only 33% among those receiving the donor vector alone (P < 0.05 using the log-rank test). Only postweaning mice are shown, because sex cannot be determined prior to weaning.

To investigate whether AAV2/8-ZFN coadministration also improved the magnitude of therapeutic effect at a late time point, blood glucose, hepatic G6Pase activity, and hepatic glycogen content were analyzed. These are standard means of assaying the efficacy of treatment for GSD Ia. We found that there was no significant difference between the two treatment groups when comparing their plasma glucose after 2 hours of fasting at 5 months of age (Figure 5a). However, both vector-treated groups had significantly elevated glucose during fasting, in comparison with untreated 2-week-old G6pc-/- mice. Liver G6Pase activity for either treatment group at 8 months of age did not differ from 2-week-old untreated affected controls, indicating that the G6Pase activity at 8 months was below the threshold of detection for the G6Pase activity assay (Figure 5b). Hepatic glycogen content at 8 months of age corroborated the G6Pase data, showing no significant difference between the two treatment groups (Figure 5c). However, both vector treated groups had significantly lower glycogen content in liver, in comparison with untreated 2-week-old G6pc-/- mice. Finally, quantitative reverse transcription PCR confirmed that the vector-treated groups did not differ in human G6Pase mRNA expression at 8 months of age, but the signal for each vector-treated group was highly elevated above the background observed for untreated 2-week-old knockout mice (Figure 5d). While it is therefore clear that treatment with AAV2/8-RoG6P has therapeutic biochemical effects with or without the AAV2/8-ZFN vector, we were not able to show that the AAV2/8-ZFN vector improves biochemical measures of GSD Ia. The lack of difference was likely due to the high mortality observed in the group that did not receive AAV2/8-ZFN, leading to selection for mice that responded especially well to gene therapy with AAV2/8-RoG6P alone.

Figure 5.

Figure 5

Open in a new tab

Markers of therapeutic effect show no difference between treatment groups at 8 months of age. (a) Blood glucose at 5 months following treatment. Both treatment groups showed similar blood glucose. Both were significantly different than previously published values for untreated G6pc-/- knockout mice.4 Two-week-old untreated mice were used because knockout mice do not survive past 3 weeks. (b) Hepatic G6Pase activity at 8 months of age in G6pc-/- knockout mice. No difference was observed between the two treatment groups or between either group and untreated knockout mice. Six-month-old WT mice are shown for positive reference. (c) Hepatic glycogen accumulation in treated and untreated knockout mice, and wild-type mice. Both treatment groups had comparable glycogen accumulation in their livers, though both groups had significantly more glycogen than wild-type controls. Glycogen in treated mice was also reduced compared with historical data on hepatic glycogen in knockout mice.8 An untreated control used to verify day-to-day experimental consistency was consistent with historical results. (d) Relative human G6Pase expression (vector-specific) normalized to mouse β-actin, as determined by RT-qPCR, was equivalent whether or not mice received AAV2/8-ZFN. However, both groups showed highly elevated expression compared with the background signal for untreated mice. Error bars: mean ± SD.

To overcome the problem posed by the high mortality rate leading to low sample sizes selected for similar potency, a short-term experiment was performed in which mice received the AAV2/9 vectors, because an AAV2/9 vector encoding G6Pase achieved higher biochemical correction than the equivalent AAV2/8 vector in G6pc-/- mice, and enhancing the effect was likely to enhance any differences between the two groups.8 Furthermore, the ratio of donor to ZFN vector was adjusted to improve the efficiency of gene editing.19 Mice were assayed at 3 months of age, rather than 8 months, to avoid the loss of data from longer-term therapeutic decline and mortality.

Plasma glucose after 8 hours of fasting was not significantly different between the two vector-treated groups at 12 weeks (Figure 6a). However, significant differences were observed in both hepatic G6Pase activity (P < 0.03) (Figure 6b) and hepatic glycogen content (P < 0.02) (Figure 6c) between single- and dual-vector treated groups at 3 months of age. This confirmed that coadministration of AAV2/9-ZFN with AAV2/9-RoG6P improved the therapeutic efficacy of the human G6Pase transgene.

Figure 6.

Figure 6

Open in a new tab

Administration of AAV2/9-ZFN enhanced biochemical correction at 3 months of age. Groups of mice with GSD Ia were treated with AAV2/9-RoG6P+ AAV2/9-ZFN (n = 11), or AAV2/9-RoG6P alone (n = 7). (a) Blood glucose at 3 months of age following an 8-hour fast. No difference was observed between the two groups in this assay, but both were improved compared with untreated knockout mice. Groups of “no vector” G6pc-/- mice (n = 7); and untreated wild-type mice (WT; n = 4) were controls. (b) G6Pase activity in the liver was significantly higher in dual-vector-treated mice, in comparison with either “no vector” G6pc-/- mice or single-vector-treated mice. Groups of “no vector” G6pc-/- mice (n = 3); and untreated wild-type mice (WT; n = 6) were controls. (c) Hepatic glycogen accumulation was reduced in in dual-vector treated mice, in comparison with either “no vector” G6pc-/- mice or single-vector-treated mice. No significant difference was observed between double-vector mice and wild-type mice. Error bars: mean ± SD.

Furthermore, livers from mice at both 8 and 3 months of age were sectioned and subjected to histochemical staining for G6Pase activity. Consistent with the data from biochemical assays for 8-month-old mice, staining for G6Pase showed no benefit from AAV2/8-ZFN administration in those mice (not shown). However, 3-month-old mice showed marked improvement in the dual-vector-treated group, in comparison with the single-vector group (Figure 7). Quantifying the number of G6Pase-expressing cells visible demonstrated a significant difference between the two groups, indicating that ZFN activity increased the number of cells with detectable G6Pase activity.

Figure 7.

Figure 7

Open in a new tab

G6Pase histochemical staining in the livers of 3-month-old mice reveals an improvement in G6Pase-expressing cells in mice receiving dual-vector treatment. (a) Representative G6Pase staining of an untreated knockout mouse at 10 days of age. Note the absence of expression throughout the liver. (b) Representative staining section of a wild-type mouse liver. Note the uniform brown stain throughout all hepatocytes. (c,d) Representative sections from mice that received the AAV2/9-RoG6P donor vector only, or the donor vector as well as AAV2/9-ZFN, respectively. (e) G6Pase-positive cell counts from both treatment groups (n = 7 each), demonstrating a significant enhancement in positive cell counts when both vectors were administered (P < 0.03). Error bars: mean ± SD.

To further investigate whether AAV2/8-ZFN administration could improve the longevity of donor transgene activity, a small-scale experiment was performed using AAV2/8-RoGFP administered to wild-type mice. The RoGFP vector genome resembles RoG6P, but rather than providing G6Pase under control of the minimal human G6Pase promoter, it provided eGFP under control of the LSP used for the ZFN vector as detailed in Materials and Methods. We observed almost complete loss of eGFP expression between 1 and 6 months of age unless AAV2/8-ZFN was coadministered, in which case eGFP expression persisted significantly better (Figure 8).

Figure 8.

Figure 8

Open in a new tab

Transduction of mouse Liver with AAV2/8-RoGFP. (a) Liver section one month following administration of AAV2/8-RoGFP or PBS. (b) Three representative liver sections 6 months following AAV2/8-RoGFP administration. (c) Three representative liver sections 6 months following AAV2/8-RoGFP and AAV2/8-ZFN administration. (d) Quantification of GFP+ cells per high magnification field using a Zeiss LSM 510 inverted confocal microscope; GFP+ cells counted from two representative fields of view (10× objective) of four mouse livers per each group and average number of cells calculated. Magnification 100×; bar is 100 µm. Mice received 1E13 vp/g IP at 2 weeks of age. Liver sections are 30 µm. Error bars: mean ± SD.

Discussion

Treatment of GSD Ia currently entails a lifelong dietary regimen designed to prevent hypoglycemia, the primary concern resulting in early death.26 This approach greatly extends the lifespan of patients, but fails to prevent many of the long-term complications of the disease like hepatic adenomas that may progress to hepatocellular carcinoma, and progressive renal failure. These complications likely result from downstream metabolic anomalies caused by the inability to release hepatic glucose. As such, approaches to improving disease prognosis as well as quality of life must target the root cause of the disease.

AAV-based gene therapies seek to treat GSD Ia at this root by supplementing the missing human G6Pase gene with a therapeutic transgene encoding G6Pase. However, targeting the developing liver of young patients would require multiple doses due to rapidly-expanding hepatocytes diluting out the episomal transgene delivered by AAV.10 As a result of this, multiple repeat doses are required to maintain a therapeutic effect over the course of many years.17 Since each dose requires a different AAV capsid to evade a recipient's antibodies against previous vectors, and each capsid preferentially targets different tissues, it is currently infeasible to treat a human patient in this manner. In order to overcome this temporal limitation, a permanent therapy for GSD Ia is required.

Here, we describe a gene therapy approach using two AAV vectors, one encoding a ROSA26-targeting ZFN, and the other encoding human G6Pase with homologous donor arms for ROSA26 surrounding the transgene. This dual-vector system is designed to cleave the mouse ROSA26 safe harbor locus to induce homology-directed repair at this site so the donor vector is copied into the mouse genome. Once in the mouse genome, the transgene should provide a lifelong source of G6Pase activity from that hepatocyte and its subsequent daughter cells. Though this is expected to occur in only a small subset of hepatocytes, it is known that as little as 3% of wild-type G6Pase activity can prevent the symptoms of hypoglycemia.13 Of note, hepatic correction with gene editing will not reverse the progressive renal involvement in GSD Ia, and further development of methods to correct G6Pase deficiency in the kidney is needed.

We show that the ROSA26-targeting ZFN gene is active in hepatocytes when delivered by AAV, ensuring that DNA damage-induced repair mechanisms are triggered to enable G6Pase transgene integration. Moreover, we observed targeted integration events dependent on AAV2/8-ZFN administration, confirming that the ZFN does induce integration as hypothesized. Achieving integration of a fully-functional transgene at a safe harbor locus in vivo provides proof-of-principle for future work on similar applications not only for GSD Ia, but in a number of other monogenic diseases as well. That ability to apply safe harbor loci-targeting ZFNs to multiple therapies is what makes this approach particularly valuable.

Prior to this work, an in vivo AAV-delivered ZFN therapy was found effective in a mouse model of hemophilia B, and used a ZFN specific to the deficient blood coagulation factor IX locus.19,20 This approach to genome-editing requires the development and refinement of nucleases specific to each therapeutically relevant gene, as well as downstream animal models and clinical trials for each. Our approach would circumvent at least part of the process by establishing a universal safe harbor nuclease AAV vector that could be paired with various therapeutic transgenes. In particular, we anticipate that this approach could be developed for human therapies by targeting the AAVS1 safe harbor locus, which has already been used similarly in vitro.27,28,29 Another potential target for AAV-based human monogenic gene therapy is the albumin locus, which is currently being investigated in murine models.24,30

Blood glucose following fasting, hepatic G6Pase activity, and hepatic glycogen content are standard metrics of gene therapy efficacy in GSD Ia mice.9 G6Pase activity for both mouse groups is below the threshold of detection by 8 months, but their hepatic glycogen content was improved (reduced) compared with untreated mice. The reduced glycogen content indicates that G6Pase activity must have taken place prior to tissue collection, and was strong enough to prevent life-threatening glycogen accumulation. A shorter time course confirmed that at 3 months of age, mice that received the ZFN vector had greater G6Pase activity and reduced glycogen, in comparison with mice that did not receive the ZFN vector. Similarly, qPCR shows that both groups contain similar levels of the transgene in their livers by 3 months, with the trend holding at 8 months. Furthermore, RT-qPCR demonstrated similar expression levels of the transgene between both groups at 8 months, highly elevated above background levels in untreated mice. Most importantly as an overall metric of therapeutic effect, we observe a significant increase in lifespan when mice were given AAV2/8-ZFN, bringing the 8-month survival rate up from 43 to 100%. The trend was also observed in female mice alone, bringing their survival from 33 to 100%. Since females are known to take up AAV vectors less efficiently into hepatocytes than males, leading to decreased gene therapy efficaciousness, the striking improvement observed in females is noteworthy.25 Furthermore, mice highly expressing the ZFN early in life showed no signs of mortality specific to receipt of AAV-ZFN, which argued against toxicity induced by the ZFN's presence.

An interesting phenomenon also emerged that was not initially predicted: an increase in observed allele modification in knockout mice compared with their normal littermates (Figure 1a), which we believe occurred due to a selective advantage for stably transduced G6Pase-containing hepatocytes in GSD Ia. More vector-containing hepatocytes were present in ZFN-treated mice, based upon increased G6Pase staining in dual-vector mice in comparison with single-vector treated mice at 3 months of age (Figure 7). A subpopulation of those G6Pase-positive cells had integrated the vector due to ZFN activity in dual-vector-treated mice. The same ZFN activity that leads to allele modification therefore also correlates with an increased likelihood of a cell stably expressing G6pase. It is possible that a selective advantage for G6Pase-positive cells would be present in the GSD Ia liver due to the cytotoxic effects of G6Pase deficiency, as postulated by Grompe and others.31 Such a selective advantage would be absent in the wild-type mouse liver. Therefore, we would expect that a selective advantage in the GSD Ia liver for G6Pase-positive hepatocytes that contained the ZFN could explain the higher rate of allele modification in G6pc-/- mice, in comparison with their normal littermates.

The low survival of non-ZFN-treated mice could explain why biochemical differences between the two treatment groups were undetectable by 8 months of age. The high mortality in mice that did not receive AAV2/8-ZFN selected for mice that responded strongly to gene therapy by 8 months of age, while those mice with a poor therapeutic response died prior to tissue collection and analysis. This is an inevitable outcome for the disease model due to near-universal mortality by 2–3 weeks of age in absence of gene therapy. We overcame this limitation by improving the ZFN treatment and performing an abbreviated experiment to ensure data collection occurred prior to late-onset mortality, and we observed enhanced biochemical correction of the liver in mice that received the dual-vector treatment, in comparison with those mice that received the donor vector alone. In conclusion, this study justifies the further development of gene editing as an approach to curative therapy for GSD Ia.

Materials and Methods

Preparation of AAV vectors. The AAV vector plasmid, pAAV-RoG6P, contained the vector genome comprised of a terminal repeat at each end flanking a transgene comprised of the human G6Pase minimal promoter to drive a human G6Pase cDNA followed by a human growth hormone polyadenylation signal, which was flanked by sequences from exon 1 of the mouse ROSA26 locus.32 The G6Pase-encoding transgene was previously described.4 The AAV vector plasmid, pAAV-ZFN, contained the transgene for the two subunits of the ROSA26-targeting “R4L6 eZFN”21 separated by a T2A self-cleavage peptide and expressed from the LSP and flanked by terminal repeats. The LSP in AAV2/8-ZFN and AAV2/9-ZFN contained a thyroid hormone-binding globulin promoter sequence downstream from two copies of an α1-microglobulin/bikunin enhancer sequence, and previously achieved long-term efficacy in hemophilia B mice within an AAV vector encoding coagulation factor IX.22 The AAV vector plasmid, pAAV-RoGFP, resembled pAAV-RoG6P except it contained eGFP expressed by the same LSP used for pAAV-ZFN, instead of a human G6Pase promoter driving G6Pase expression. The AAV vectors were pseudotyped with AAV capsid proteins as described (packaging plasmids courtesy of Dr. James M. Wilson, University of Pennsylvania, Philadelphia, PA), and the helper plasmid was pAdHelper (Stratagene, La Jolla, CA).4

AAV vector administration to G6pc-/- mice. Carrier G6Pase (+/-) mice were housed in the Duke Vivarium and bred to produce homozygous, affected G6pc-/- offspring. Affected genotype was confirmed by PCR analysis of tail DNA with primers within and flanking the neo gene insertion in the G6Pase gene as described.9 Heterozygous G6pc+/- mice were pooled with homozygous wild-type mice throughout the experiments, as G6pc+/- heterozygotes are biochemically equivalent to wild-type mice and are routinely treated as such in the literature.12,33 G6pc-/- mice were injected with vectors via the retro-orbital sinus at 12 ± 1 days of age without regard to sex, and both males and females were included in all groups. Injection was performed following isoflurane anesthesia with a 28 gauge insulin syringe (10 μl/g volume; 2E13 VP/kg each vector for mice later collected at 8 months of age; 1.3E13 VP/kg AAV2/9-RoG6P and 4.8E12 VP/kg AAV2/9-ZFN for mice later collected at 3 months of age), and hemostasis was achieved by brief manual pressure. Daily injection of 0.1 to 0.2 ml 10% dextrose subcutaneously was initiated at 3 days of age and continued for 2–3 weeks. Mice used in the 3-month time course were given such dextrose injections twice-daily for 3–4 weeks. Mice were fasted periodically for 2 hours beginning at 10 am, or for 8 hours beginning at 8 am. All mouse procedures were done in accordance with Duke University Institutional Animal Care and Use Committee-approved guidelines.

Quantification of DNA repair at the ROSA26 locus in the liver. Liver DNA was extracted using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI). The ROSA26 locus was PCR amplified by Takara ExTaq (Takara Clontech, Mountain View, CA) with the following reagents: 2.5 μl ExTaq buffer; 2 μl 2.5 mmol/l dNTP mix; 1 μl 10 μmol/l primer SrvF1 (5′-AAGGGAGCTGCAGTGGAGTA-3′)21; 1 μl 10 μmol/l primer SrvR1 (5′-GCGGGAGAAATGGATATGAA-3′)21; 17.3 μl water, 1 μl (100 ng) genomic DNA; 0.2 μl HotStart ExTaq polymerase. Cycling conditions were: 20 cycles of melting at 98° for 10 seconds, annealing at temperatures reduced each cycle by 0.5° from 70-60.5° for 10 seconds, extension at 72° for 30 seconds; then 20 cycles of melting at 98° for 10 seconds, annealing at 60° for 10 seconds, extension at 72° for 30 seconds; and holding at 4°. First-round PCR products were diluted 1:100 and 1 μl used in a nested reaction with the same conditions except primers were SrvFnst (5′-GGGAGGTGTGGGAGGTTT-3′) and SrvRnst (5′-TGGCCACTCGTTTAAACCTC-3′). Second-round PCR products were self-hybridized by incubation in a thermocycler with the following conditions at a −0.1°/second ramp rate: 95° for 3 minutes; 85° for 20 seconds; 75° for 20 seconds; 65° for 20 seconds; 55° for 20 seconds; 45° for 20 seconds; 35° for 20 seconds; 25° for 20 seconds; hold at 4°. 17.8 μl of the hybridized second-round products were then incubated with 2.2 μl of 150 mmol/l MgCl2, 1 μl of Surveyor nuclease (Transgenomic, Omaha, NE), and 1 μl of Surveyor Enhancer S at 42° for 1 hour prior to loading on a 10% PAGE-TBE gel. The gel was stained with ethidium bromide and analyzed with densitometry to quantify the prevalence of NHEJ DNA repair at the ROSA26 locus.

Identification of transgene/mouse-genome junctions. Genomic DNA was extracted from mouse liver as described above. The first round of PCR was set up as follows: 500 ng genomic DNA; 2.5 μl of 10× standard Taq buffer with 15 mmol/l MgCl2 (Qiagen, Venlo, Limburg); 1 μl of 25 mmol/l MgCl2; 3 μl of 2.5 mmol/l dNTP mix; Q-Solution (Qiagen), 1.25 μl of 10 μmol/l Primer 1 (5′-GTAATCAATACCATGTGGCTC-3′); 1.25 μl of 10 μmol/l Primer 2 (5′-TCGAGCTGGTCTTCTACGTC-3′); 0.25 μl Taq (Qiagen); and water to 25 μl. Cycling conditions were: 95° for 3 minutes; then 35 cycles of 95° for 30 seconds, 58.1° for 30 seconds, 72° for 2 minutes 20 seconds; then 72° for 5 minutes and hold at 4°. 2 μl of a 1:200 dilution of the first-round PCR was used in the second round, which was the same as the first round save that extra MgCl2 was not added, the primers were Primer 3 (5′-GACATCCACCTGGAAACCATT-3′) and Primer 4 (5′-CGTCAGTGTCATCCCCTACT-3′), and annealing was done at 54.1° instead of 58.1°. Second-round PCR products were subcloned using a PCR cloning kit (Qiagen) per the manufacturer's instructions. Resulting plasmids were grown in XL1-Blue Supercompetent Cells (Agilent Technologies, Santa Clara, CA) and Sanger sequenced with the T7 Promoter and SP6 primers from Eton Bioscience (San Diego, CA).

Quantification of vector DNA in the liver. Quantitative real-time PCR was performed using SYBR green in a LightCycler 480II (Roche, Basel, Switzerland) following the manufacturer's instructions. Gene-specific primers for the human G6Pase promoter (sense 5′-CAAAGATCAGGGCTGGGTTGA-3′, and antisense 5′-CTTGGTGGTGATTGCTCTGCT-3′), and for mouse β-actin (sense 5′-AGAGGGAAATCGTGCGTGAC-3′ and antisense 5′-CAATAGTGATGACCTGGCCGT-3′) were used for each reaction. Plasmid DNA corresponding to 0.01 copy to 100 copies of human G6Pase gene (in 500 ng genomic DNA) was used in a standard curve. To determine the vg copy number, the Cp values of samples were compared to the standard curve. Cycling conditions were 5 minutes at 95 °C, followed by 45 cycles of 95 °C for 10 seconds, 60 °C for 10 seconds, and 72 °C for 20 seconds followed by acquisition.

Evaluation of biochemical correction. Enzyme analysis was performed as described.4 Briefly, tissues were frozen and stored at −80 °C. Glycogen content was measured by complete digestion of polysaccharide using amyloglucosidase (Sigma Chemical, St. Louis, MO). The structure of the polysaccharide was inferred by using phosphorylase free of the debranching enzyme to measure the yield of glucose-1-phosphate. Specific G6Pase activity was measured by using glucose-6-phosphate as substrate after subtraction of nonspecific phosphatase activity as estimated by β-glycerophosphate. Glucose was analyzed with a Kodak Biolyzer (Eastman Kodak Company, Rochester, NY) according to the manufacturer's recommendations.

Histochemical staining for G6Pase activity in the liver. G6Pase was detected qualitatively in frozen sections (6 μm) mouse liver by an optimized cerium-diaminobenzidine method as described.8 G6Pase-expressing cells were counted from ten 20× fields were counted for each mouse and averaged, then counts were averaged across treatment groups.

Quantification of human G6Pase mRNA expression in the liver. Quantitative real-time PCR was performed on cDNA reverse transcribed from total RNA collected from mouse liver tissue. RNA purification was done using Trizol (Invitrogen, Waltham, MA) following the manufacturer's instructions, and RNA was dissolved in 25 μl of water. Ten microliters of RNA were treated to remove DNA using the DNA-free DNA Removal Kit (Ambion, Waltham, MA) in half-reactions of the manufacturer's routine DNase treatment instructions, with an extended DNase treatment time of 1 hour. RevertAid Reverse Transcriptase (Thermo Scientific, Waltham, MA) was used with random hexamers to generate cDNA from 5uL of DNA-free RNA, per the manufacturer's instructions. Quantitative real-time PCR was performed using SYBR green (Qiagen) in a LightCycler 480II (Roche, Basel, Switzerland) following the manufacturer's instructions. Input cDNA was 1 μl of cDNA diluted 1:10 with water. Gene-specific primers for human G6Pase cDNA (sense 5′-CTGTTCAGCTTCGCCATC-3′, and antisense 5′-GGGAGGCTACAATAGAGCT-3′), and for mouse β-actin (sense 5′-AGAGGGAAATCGTGCGTGAC-3′ and antisense 5′-CAATAGTGATGACCTGGCCGT-3′) were used for each reaction. Cycling conditions were 5 minutes at 95 °C, followed by 45 cycles of 95 °C for 10 seconds, 60 °C for 10 seconds, and 72 °C for 20 seconds followed by acquisition. Relative expression was calculated using the ΔΔCt method.34

AAV vector administration and eGFP quantification in wild-type and carrier mice. Wild-type and carrier mice used for the eGFP experiment were bred as described in Materials and Methods. They were not given subcutaneous dextrose, and vector administration was intraperitoneal (10 μl/g volume of AAV2/8-ZFN for 1E13 VP/kg, and 5 μl/g volume of AAV2/8-RoGFP for 1E13 VP/kg), which has been shown to have comparable transduction to veinous routes.35

Mouse livers were collected at 1 or 6 months of age and fixed in 4% paraformaldehyde overnight. Tissue was then transferred to 30% sucrose for 2 days prior to being flash-frozen in OCT and stored at −80 °C. Frozen 30 µm sections were imaged for eGFP fluorescence on a Zeiss LSM 510 inverted confocal microscope at 10× magnification and eGFP-positive cells were quantified by counting bright cells from two representative sections per mouse.

Statistical analysis. The heteroscedastic two-tailed T-test was used to compare groups unless otherwise noted. Survival analysis included production of Kaplan-Meier curves. P values of less than 0.05 were considered significant.

Acknowledgments

This work was funded by NIH grant R01DK105434 from the National Institute of Diabetes and Digestive and Kidney Diseases. Funding was provided by the Alice and YT Chen Center for Pediatric Genetics and Genomics, The Love of Christopher fund and from John Kelly. The authors have no conflicts of interest to report with regard to this work.

References

  1. Koeberl, DD, Kishnani, PS, Bali, D and Chen, YT (2009). Emerging therapies for glycogen storage disease type I. Trends Endocrinol Metab 20: 252–258. [DOI] [PubMed] [Google Scholar]
  2. Ghosh, A, Allamarvdasht, M, Pan, CJ, Sun, MS, Mansfield, BC, Byrne, BJ et al. (2006). Long-term correction of murine glycogen storage disease type Ia by recombinant adeno-associated virus-1-mediated gene transfer. Gene Ther 13: 321–329. [DOI] [PubMed] [Google Scholar]
  3. Jiang, H, Lillicrap, D, Patarroyo-White, S, Liu, T, Qian, X, Scallan, CD et al. (2006). Multiyear therapeutic benefit of AAV serotypes 2, 6, and 8 delivering factor VIII to hemophilia A mice and dogs. Blood 108: 107–115. [DOI] [PubMed] [Google Scholar]
  4. Koeberl, DD, Pinto, C, Sun, B, Li, S, Kozink, DM, Benjamin, DK Jr et al. (2008). AAV vector-mediated reversal of hypoglycemia in canine and murine glycogen storage disease type Ia. Mol Ther 16: 665–672. [DOI] [PubMed] [Google Scholar]
  5. Sarkar, R, Mucci, M, Addya, S, Tetreault, R, Bellinger, DA, Nichols, TC et al. (2006). Long-term efficacy of adeno-associated virus serotypes 8 and 9 in hemophilia a dogs and mice. Hum Gene Ther 17: 427–439. [DOI] [PubMed] [Google Scholar]
  6. Seppen, J, Bakker, C, de Jong, B, Kunne, C, van den Oever, K, Vandenberghe, K et al. (2006). Adeno-associated virus vector serotypes mediate sustained correction of bilirubin UDP glucuronosyltransferase deficiency in rats. Mol Ther 13: 1085–1092. [DOI] [PubMed] [Google Scholar]
  7. Zincarelli, C, Soltys, S, Rengo, G and Rabinowitz, JE (2008). Analysis of AAV serotypes 1-9 mediated gene expression and tropism in mice after systemic injection. Mol Ther 16: 1073–1080. [DOI] [PubMed] [Google Scholar]
  8. Luo, X, Hall, G, Li, S, Bird, A, Lavin, PJ, Winn, MP et al. (2011). Hepatorenal correction in murine glycogen storage disease type I with a double-stranded adeno-associated virus vector. Mol Ther 19: 1961–1970. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Koeberl, DD, Sun, BD, Damodaran, TV, Brown, T, Millington, DS, Benjamin, DK Jr et al. (2006). Early, sustained efficacy of adeno-associated virus vector-mediated gene therapy in glycogen storage disease type Ia. Gene Ther 13: 1281–1289. [DOI] [PubMed] [Google Scholar]
  10. Cunningham, SC, Dane, AP, Spinoulas, A, Logan, GJ and Alexander, IE (2008). Gene delivery to the juvenile mouse liver using AAV2/8 vectors. Mol Ther 16: 1081–1088. [DOI] [PubMed] [Google Scholar]
  11. Weinstein, DA, Correia, CE, Conlon, T, Specht, A, Verstegen, J, Onclin-Verstegen, K et al. (2010). Adeno-associated virus-mediated correction of a canine model of glycogen storage disease type Ia. Hum Gene Ther 21: 903–910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Yiu, WH, Lee, YM, Peng, WT, Pan, CJ, Mead, PA, Mansfield, BC et al. (2010). Complete normalization of hepatic G6PC deficiency in murine glycogen storage disease type Ia using gene therapy. Mol Ther 18: 1076–1084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Lee, YM, Jun, HS, Pan, CJ, Lin, SR, Wilson, LH, Mansfield, BC et al. (2012). Prevention of hepatocellular adenoma and correction of metabolic abnormalities in murine glycogen storage disease type Ia by gene therapy. Hepatology 56: 1719–1729. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Sun, B, Li, S, Yang, L, Damodaran, T, Desai, D, Diehl, AM et al. (2009). Activation of glycolysis and apoptosis in glycogen storage disease type Ia. Mol Genet Metab 97: 267–271. [DOI] [PubMed] [Google Scholar]
  15. Koeberl, DD, Sun, B, Bird, A, Chen, YT, Oka, K and Chan, L (2007). Efficacy of helper-dependent adenovirus vector-mediated gene therapy in murine glycogen storage disease type Ia. Mol Ther 15: 1253–1258. [DOI] [PubMed] [Google Scholar]
  16. Raju, BI, Leyvi, E, Seip, R, Sethuraman, S, Luo, X, Bird, A et al. (2013). Enhanced gene expression of systemically administered plasmid DNA in the liver with therapeutic ultrasound and microbubbles. IEEE Trans Ultrason Ferroelectr Freq Control 60: 88–96. [DOI] [PubMed] [Google Scholar]
  17. Demaster, A, Luo, X, Curtis, S, Williams, KD, Landau, DJ, Drake, EJ et al. (2012). Long-term efficacy following readministration of an adeno-associated virus vector in dogs with glycogen storage disease type Ia. Hum Gene Ther 23: 407–418. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Perez-Pinera, P, Ousterout, DG and Gersbach, CA (2012). Advances in targeted genome editing. Curr Opin Chem Biol 16: 268–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Li, H, Haurigot, V, Doyon, Y, Li, T, Wong, SY, Bhagwat, AS et al. (2011). In vivo genome editing restores haemostasis in a mouse model of haemophilia. Nature 475: 217–221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Anguela, XM, Sharma, R, Doyon, Y, Miller, JC, Li, H, Haurigot, V et al. (2013). Robust ZFN-mediated genome editing in adult hemophilic mice. Blood 122: 3283–3287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Perez-Pinera, P, Ousterout, DG, Brown, MT and Gersbach, CA (2012). Gene targeting to the ROSA26 locus directed by engineered zinc finger nucleases. Nucleic Acids Res 40: 3741–3752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Wang, L, Takabe, K, Bidlingmaier, SM, Ill, CR and Verma, IM (1999). Sustained correction of bleeding disorder in hemophilia B mice by gene therapy. Proc Natl Acad Sci USA 96: 3906–3910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Guschin, DY, Waite, AJ, Katibah, GE, Miller, JC, Holmes, MC and Rebar, EJ (2010). A rapid and general assay for monitoring endogenous gene modification. Methods Mol Biol 649: 247–256. [DOI] [PubMed] [Google Scholar]
  24. Barzel, A, Paulk, NK, Shi, Y, Huang, Y, Chu, K, Zhang, F et al. (2015). Promoterless gene targeting without nucleases ameliorates haemophilia B in mice. Nature 517: 360–364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Davidoff, AM, Ng, CY, Zhou, J, Spence, Y and Nathwani, AC (2003). Sex significantly influences transduction of murine liver by recombinant adeno-associated viral vectors through an androgen-dependent pathway. Blood 102: 480–488. [DOI] [PubMed] [Google Scholar]
  26. Kishnani, PS, Austin, SL, Abdenur, JE, Arn, P, Bali, DS, Boney, A et al.; American College of Medical Genetics and Genomics. (2014). Diagnosis and management of glycogen storage disease type I: a practice guideline of the American College of Medical Genetics and Genomics. Genet Med 16: e1. [DOI] [PubMed] [Google Scholar]
  27. Hockemeyer, D, Soldner, F, Beard, C, Gao, Q, Mitalipova, M, DeKelver, RC et al. (2009). Efficient targeting of expressed and silent genes in human ESCs and iPSCs using zinc-finger nucleases. Nat Biotechnol 27: 851–857. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. DeKelver, RC, Choi, VM, Moehle, EA, Paschon, DE, Hockemeyer, D, Meijsing, SH et al. (2010). Functional genomics, proteomics, and regulatory DNA analysis in isogenic settings using zinc finger nuclease-driven transgenesis into a safe harbor locus in the human genome. Genome Res 20: 1133–1142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Tay, FC, Tan, WK, Goh, SL, Ramachandra, CJ, Lau, CH, Zhu, H et al. (2013). Targeted transgene insertion into the AAVS1 locus driven by baculoviral vector-mediated zinc finger nuclease expression in human-induced pluripotent stem cells. J Gene Med 15: 384–395. [DOI] [PubMed] [Google Scholar]
  30. Sharma, R, Anguela, XM, Doyon, Y, Wong, SY, Paschon, DE, Li, H et al. In: Vivo ZFN Mediated Targeting of Albumin as a Platform for Expression of Multiple Therapeutic Genes. The American Society of Gene and Cell Therapy, vol. 21. Molecular Therapy: Salt Lake City, Utah, 2013. pp. S188–189. [Google Scholar]
  31. Grompe, M (2006). Principles of therapeutic liver repopulation. J Inherit Metab Dis 29: 421–425. [DOI] [PubMed] [Google Scholar]
  32. Mural, RJ, Adams, MD, Myers, EW, Smith, HO, Miklos, GL, Wides, R et al. (2002). A comparison of whole-genome shotgun-derived mouse chromosome 16 and the human genome. Science 296: 1661–1671. [DOI] [PubMed] [Google Scholar]
  33. Lei, KJ, Chen, H, Pan, CJ, Ward, JM, Mosinger, B Jr, Lee, EJ et al. (1996). Glucose-6-phosphatase dependent substrate transport in the glycogen storage disease type-1a mouse. Nat Genet 13: 203–209. [DOI] [PubMed] [Google Scholar]
  34. Livak, KJ and Schmittgen, TD (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25: 402–408. [DOI] [PubMed] [Google Scholar]
  35. Dane, AP, Wowro, SJ, Cunningham, SC and Alexander, IE (2013). Comparison of gene transfer to the murine liver following intraperitoneal and intraportal delivery of hepatotropic AAV pseudo-serotypes. Gene Ther 20: 460–464. [DOI] [PubMed] [Google Scholar]

Articles from Molecular Therapy are provided here courtesy of The American Society of Gene & Cell Therapy

RESOURCES