Long-term rescue of a lethal murine model of methylmalonic acidemia using adeno-associated viral gene therapy.

Methylmalonic acidemia (MMA) is an organic acidemia caused by deficient activity of the mitochondrial enzyme methylmalonyl-CoA mutase (MUT). This disorder is associated with lethal metabolic instability and carries a poor prognosis for long-term survival. A murine model of MMA that replicates a severe clinical phenotype was used to examine the efficacy of recombinant adeno-associated virus (rAAV) serotype 8 gene therapy as a treatment for MMA. Lifespan extension, body weight, circulating metabolites, transgene expression, and whole animal propionate oxidation were examined as outcome parameters after gene therapy. One-hundred percent of the untreated Mut(-/-) mice (n = 58) died by day of life (DOL) 72, whereas >95% of the adeno-associated virus-treated Mut(-/-) mice (n = 27) have survived for > or = 1 year. Despite a gradual loss of transgene expression and elevated circulating metabolites in the treated Mut(-/-) mice, the animals are indistinguishable from unaffected control littermates in size and activity levels. These experiments provide the first definitive evidence that gene therapy will have clinical utility in the treatment of MMA and support the development of gene therapy for other organic acidemias.

Introduction

Methylmalonic acidemia (MMA) is a severe autosomal recessive inborn error of intermediary metabolism characterized by intermittent metabolic instability, multiorgan pathology, growth retardation, and a poor prognosis for long-term survival.[1,2,3,4,5,6] The disorder exhibits genetic heterogeneity and can be caused by deficient enzymatic activity of methylmalonyl-CoA mutase (MUT) or defective intracellular transport, processing, and metabolism of cobalamin.[7] MUT is an important mitochondrial enzyme in propionyl-CoA metabolism and converts L-methylmalonyl-CoA into succinyl-CoA, a Krebs cycle intermediate. A block at this enzymatic step results in elevated plasma levels of methylmalonic acid as well the accumulation of other propionyl-CoA-derived metabolites such as 2-methylcitrate.[8] The etiology of the many medical problems that patients with MMA suffer is not well understood. However, the wide spectrum of severity that can be seen in patients who harbor missense mutations suggests that restoring a very low level of enzyme activity would provide substantial clinical benefit.

Currently, the main treatment for affected patients is dietary restriction of propiogenic amino acids to reduce circulating metabolites. Liver[9,10,11,12,13,14,15] and/or combined liver/kidney[16,17] transplantation has been performed in a limited fashion in an attempt to improve metabolic stability through the provision of organ-specific enzymatic activity. Although this approach has been effective, and even curative, for other metabolic disorders,[18] the clinical utility of solid organ transplantation as a standard treatment for MMA is unclear given the small number of patients that have undergone the procedure.[14,15,19] The need for new and widely available therapies for MMA is underscored by a recent multicenter European long-term patient study, which described an overall 46% (n = 52) mortality for patients with MMA resulting from MUT deficiency by 30 years of age.[6]

We have previously developed a murine model of MMA that replicates the severest clinical phenotype of MMA seen in patients.[20] Affected animals display a 100–200-fold increase in plasma methylmalonic acid concentrations and most perish within the first few days of life.[20,21] Recently, we reported that direct hepatic, but not intramuscular, injection of an adenovirus that expressed the Mut gene under the control of a cytomegalovirus promoter could only partially rescue the Mut mice.[22] A liver-directed approach was initially selected because the hepatocytes of Mut mice and MMA patients manifest morphological changes and display a severe secondary electron transport chain defect that is likely contributory to the pathology of the disease.[21] Additionally, we had demonstrated that robust correction of human MUT hepatocytes was feasible with this vector.[23] However, in the surviving adenoviral-treated Mut−/− mice, the Mut levels steadily declined, and most animals died between 1 and 3 months after treatment. These results suggested that persistent expression would be needed to provide long-term amelioration of the lethal phenotype in this murine model of MMA.

Recombinant adeno-associated viruses (rAAVs) have been successfully used as gene delivery vehicles in numerous animal models of human disease and have yielded long-term transgene expression without vector-related toxicity. Furthermore, clinical trials using rAAV have involved hundreds of subjects, supporting the safety of these vectors for use in humans.[24] Over 100 different natural AAV serotypes have been isolated from a variety of species and some display striking tissue tropism.[25] Several studies have demonstrated that rAAV serotype 8 (rAAV8) vectors can transduce mouse hepatocytes with high efficiency when delivered via the portal vein or intraperitoneal injection in the neonatal period.[26,27,28] Similarly, rAAV8 vectors can efficiently transduce skeletal muscle,[29,30] a tissue that makes a significant contribution to the circulating metabolite pool in patients with MMA.[20]

In this report, we describe the therapeutic efficacy of a rAAV vector as a new gene therapy treatment for MMA and apply a novel stable isotope metabolic method to monitor the function of Mut after gene therapy. The treated Mut mice have reduced circulating metabolites, are phenotypically indistinguishable from their unaffected Mut mice littermates and show sustained enzymatic activity 1 year after treatment with the rAAV8 vector. Our results provide the first evidence that systemic gene delivery using rAAV should be useful as a treatment for patients with MMA and other organic acidemias, disorders that currently lack definitive therapy. This gene delivery platform and metabolic monitoring technique should be immediately translatable to a human gene therapy trial for MMA.

Results

Gene therapy rescues the lethal Mut−/− phenotype

Mut−/− newborn mice received a direct hepatic injection with either 1 or 2 × 1011 vector genome copies (GCs) of rAAV8-mMut. All Mut−/− mice (n = 27) injected in the neonatal period with rAAV8-mMut survived until day of life (DOL) 90 (). A single Mut−/− mouse from the rAAV8-mMut 1 × 1011 GC group perished at DOL 92 following a blood collecting procedure; a full necropsy was performed, and no abnormalities were observed. Ninety-six percent (26/27) of the Mut−/− mice treated with rAAV8-mMut have survived beyond a year. In contrast, 100% of the untreated mutants (n = 58) perished by DOL 72, with >90% of this group dying by DOL 24. In another control group, newborn Mut−/− mice (n = 18) received a direct hepatic injection with either 1 or 2 × 1011 GC of rAAV8 containing green fluorescent protein (GFP) complementary DNA driven by the same chicken β-actin promoter/enhancer (). None of the rAAV8-GFP-treated Mut−/− mice survived beyond DOL 3. Untreated Mut−/− mice and rAAV8-GFP-treated Mut−/− animals were found dead or cannibalized; the exact causes of death were undetermined. To determine whether direct hepatic injection was necessary to rescue Mut−/− mice, newborn Mut−/− mice (n = 4) received an intraperitoneal injection of 3 × 1011 GC rAAV8-mMut. Three out of the four (75%) of these mice survived and are still alive at DOL 120 (). Other than mild hepatomegaly, the Mut−/− mice (n = 4) treated by rAAV8-mMut gene therapy and sacrificed for tissue harvesting on DOL 90 or at 1 year had no gross pathologic changes noted on dissection.

Correction of growth retardation following gene therapy

Mut−/− mice were indistinguishable from their Mut+/− littermates at birth, but those rare mutants that escaped neonatal lethality were grossly abnormal. The small number of untreated Mut−/− mice that survived to DOL 24 (6/58) and 60 (3/58) were severely growth retarded, weighing <40% of sex-matched Mut+/− littermates (). Mut−/− mice treated with an intrahepatic injection of either 1 or 2 × 1011 GC of rAAV8-mMut were grossly indistinguishable in size and behavior compared to controls () and achieved body weights that were similar to their Mut+/− sex-matched littermates on DOL 24 and 60 ().

Expression of Mut after treatment with rAAV8-mMut

Mut expression in liver and muscle samples from treated Mut−/− mice at DOL 90 and 1 year of life was analyzed using quantitative PCR (qPCR) to measure mRNA levels and by western blotting to examine protein content. The Mut−/− mice at DOL 90 injected with either 1 or 2 × 1011 GC of rAAV8 expressed 38 and 72% of the endogenous Mut mRNA levels found in the liver of untreated Mut+/− animals (), whereas the level of Mut mRNA in the lower limb skeletal muscle of treated animals exceeded the endogenous Mut transcript levels measured to control heterozygotes (). Even greater expression was noted in the hearts of the treated Mut−/− mice compared to untreated Mut+/− mice, consistent with previous observations of highly efficient transduction of cardiac and skeletal myocytes by rAAV8 vectors.[30] Mut mRNA was variably detected in whole brain extracts and not detected in the kidney or spleen. Western blotting showed Mut protein in both the liver and the skeletal muscle of the treated Mut−/− mice at levels that paralleled those observed in the qPCR experiments on DOL 90 (). Mice studied at longer times showed persistent expression of the Mut transgene. The levels of Mut mRNA in the liver and skeletal muscle of treated Mut−/− mice 1 year after injection diminished but was still readily detectable by qPCR (), but not by western analysis.

Gene therapy restores Mut function and activity

Several parameters reflective of Mut enzymatic function were examined in the treated Mut−/− mice. Circulating metabolites were measured and taken to reflect whole-body Mut enzymatic activity because all the mice ingested a precursor unrestricted diet. The plasma methylmalonic acid concentrations in the treated Mut−/− mice were significantly lower than in untreated Mut−/− mice at both time points measured (). The untreated Mut−/− mice had mean plasma methylmalonic acid concentrations of 1,342 and 1,120 µmol/l on days 24 and 60. No untreated Mut−/− mice survived beyond day 72; therefore, plasma methylmalonic acid levels after day 60 from this group could not be obtained. Both the 1 and 2 × 1011 GC groups of treated Mut−/− mice had mean plasma methylmalonic acid concentrations between 440 and 540 µmol/l at the day 24 and 60 time points. Metabolites in these two groups of treated Mut−/− mice were also measured at 90, 120, 180, and 360 days with the mean methylmalonic acid levels at these time points ranging from 365 to 596 µmol/l (). As observed in humans with MMA whom have received liver or combined liver–kidney transplants, plasma metabolites were not normalized[13,16,17] and remained 50–100-fold increased over the level seen in unaffected Mut+/− mice (5–10 µmol/l).

To examine whether the long-term survival and ameliorated metabolite levels observed in the treated Mut−/− mice corresponded with increased whole-body enzyme activity, we developed a novel in vivo propionate oxidation assay. Mut−/− mice at 1 year after treatment were injected with 1-13C-sodium propionate and the subsequent metabolism of this tracer through the Mut reaction, into the Krebs cycle, with eventual oxidation into 13CO2 was determined. As can be seen in , Mut+/− mice metabolize ~70% of 1-13C-propionate into 13CO2 in 25 minutes. Untreated Mut−/− mice convert ~10% of the dose, with very flat enrichment kinetics. At 1 year of age, the treated Mut−/− mice show a markedly increased capacity to oxidize 1-13C-propionate and on average, can convert ~40% of the injected dose into 13CO2.

rAAV8-mMut rescues postneonatal Mut mice

As an extension of rAAV8-mMut gene therapy beyond the immediate neonatal period, three rare untreated Mut−/− mice that survived until DOL 20 received a single intraperitoneal injection of 3 × 1011 GC of rAAV8-mMut. At the time of injection, the animals were hypoactive, runted () and showed an impaired ability to produce 13CO2 from 1-13C-propionate. When studied 10 days after receiving the rAAV8-mMut, the mice had an improved clinical appearance (), fully restored propionate oxidation () and displayed a tenfold reduction in plasma methylmalonic acid concentrations ().

Discussion

The experiments undertaken in the present report were designed to test the efficacy of rAAV-mediated gene therapy in a murine model of MMA, a prototypical organic acidemia. Our earlier studies in Mut mice[22] and the demonstration that liver transplantation appears beneficial for a subset of MMA patients[15] led to the selection of adeno-associated virus serotype 8 as a gene delivery vector. The observed results are striking: a single intrahepatic injection of rAAV8-mMut delivered in the neonatal period was sufficient to uniformly rescue treated Mut−/− mice from certain death for over a year. The effects of gene therapy extended beyond immediate mortality and allowed the treated Mut−/− mice to gain weight, thrive, and reproduce. Limited pathological investigations have been performed on the treated mutants at older times and will be the subject of future studies, particularly to examine whether renal, hepatic, central nervous system, or pancreatic changes are present and if they have functional consequences for the treated mice. The treated Mut−/− animals were also able to tolerate a liberalized diet in the face of elevated circulating metabolites, which were greatly diminished compared to the untreated Mut−/− group, but still significantly increased compared to heterozygous controls. The Mut−/− mice, both treated and untreated, did not receive a precursor-restricted diet, commonly employed to treat patients[31] that likely would have further decreased methylmalonic acid levels in the treated animals. Complete restoration of plasma metabolites to normal in the treated mice was not expected because patients with MMA who have received replacement liver and kidney combined transplantation procedures also display persistent MMA and methylmalonic aciduria.[16,17] Also, there is no evidence to suggest that free methylmalonic acid can be efficiently metabolized, even when delivered exogenously to a wild-type host.

Consistent with many previous studies, the rAAV8-mMut vector produced persistent expression in the liver and muscle that was readily detected at 90 days after therapy at the mRNA and protein level, and at 1 year through mRNA expression and in vivo propionate oxidation. Because the mice were treated at the time of birth, the rapid growth of the liver and subsequent dilution in the number of transduced cells by cell division likely explains the relative diminution of Mut expression over time.[32] The cohort of treated Mut−/− mice, which is >25, have survived to 1 year and beyond, demonstrating that even low levels of Mut expression are sufficient to provide metabolic homeostasis, and prevent morbidity and mortality. Furthermore, although formal testing has not been performed, the mutant animals appeared clinically well, with no obvious neurological or behavioral phenotypes. rAAV-based gene vectors have previously shown promising proof-of-principle correction in other mouse models of metabolic disease[33] and now includes a pleiotropic disorder of organic acid metabolism.

The hereditary MMAs, as well as other inborn errors of metabolism that lack conventional therapy, are included in routine newborn screening panels used by many states and countries.[34] There has been a vigorous and public debate on the inclusion of these disorders in the list of conditions for which screening is offered. In this report, we have demonstrated that a single injection of rAAV8-mMut was sufficient to cure the lethal phenotype of Mut deficiency in a murine model of MMA that closely replicates the human condition. Our studies are the first to demonstrate that MMA, and by extension other organic acidemias, might be treated by gene therapy with a safe and effective vector. This conclusion offers strong support for the continued and expanded screening of infants for disorders of intermediary metabolism and to the application of gene therapy to humans with MMA.

Materials and Methods

The targeted Mut allele harbors a deletion of exon 3 in the Mut gene. This exon encodes the putative substrate-binding pocket in the Mut enzyme. The Mut allele does not produce mature RNA, protein, or enzymatic activity.[20] Mut mice on a mixed (C57BL/6 × 129SV/Ev × FvBN) background exhibit a semipenetrant neonatal lethal phenotype with most mice perishing in the early neonatal period.[21] Coat colors are variable in these mice due to parental strain contributions. Mut mice display massively elevated methylmalonic acid concentrations in the plasma that progressively rises to the 2 mmol/l range until death occurs. Mut+/− animals have biochemical parameters identical to Mut+/+ wild-type animals and were used as controls throughout.

The University of Pennsylvania Vector Core provided the expression vector, p-AAV2-CI-CB7-RBG. The vector contains transcriptional control elements from the cytomegalovirus enhancer/chicken β-actin promoter, cloning sites for the insertion of a complementary DNA, and the rabbit β-globin polyA signal.[35] Terminal repeats from AAV serotype 2 flank the expression cassette. Either the murine Mut (mMut) or GFP was cloned into pAAV2/8.CI.CB7.RBG and packaged into rAAV8, purified by cesium chloride centrifugation, and titered by qPCR as previously described.[26] pAAV2/8.CI.CB7.EGFP.RBG had a titer of 2.25 × 1013 GC/ml and pAAV2/8.CI.CB7.mMut.RBG had a titer of 4.13 × 1013 GC/ml. Animal studies were reviewed and approved by the National Human Genome Research Institute Animal User Committee. Hepatic injections were performed on nonanesthetized neonatal mice, typically within several hours after birth. Viral particles were diluted to a total volume of 20 microliters with phosphate-buffered saline immediately before injection and were delivered into the liver parenchyma using a 32-gauge needle and transdermal approach, as previously described.[22]

Total RNA was extracted using RNeasy Mini Kit (Qiagen, Valencia, CA), and DNase digested was preformed using DNA-free (Ambion, Austin, TX). qPCR was accomplished with TaqMan gene expression assays [mouse GAPD (4352932E) and murine Mut (Mm00485312_m1) from Applied Biosystems, Foster City, CA]. Samples were analyzed in an Applied Biosystems 7500 fast real-time PCR system, in accordance with the manufacturer's protocol. All samples were analyzed in triplicate. Three individual mouse tissue samples were used to determine the 100% comparator Mut Mut mRNA expression level.

Tissue samples were homogenized with a 2-ml Tenbroeck tissue grinder (Wheaton, Millville, NJ) in T-PER (Pierce Biotechnology, Rockford, IL) tissue protein extraction buffer in the presence of Halt (Pierce Biotechnology) protease inhibitor cocktail. Twenty micrograms of clarified extract were used in western analysis and probed with affinity-purified, rabbit polyclonal antisera raised against the murine Mut enzyme.[23] Complex III Core II was used as a loading control and was also detected by immunoblotting [mouse monoclonal anti-OxPhos Complex III (ubiquinol-cytochrome c oxidoreductase) Core II antibody, Invitrogen SKU# A-11143]. The antimutase antibody was used at a dilution of 1:750, and the anti-Complex III Core II antibody was used at a dilution of 1:2,000. Horseradish peroxidase–conjugated anti-rabbit IgG (NA934; GE Healthcare Life Sciences, Piscataway, NJ) or rabbit anti-goat IgG (sc-2768; Santa Cruz Biotechnology, Santa Cruz, CA) was used as the secondary antibody and was visualized with chemiluminescence detection (Pierce Biotechnology).

Plasma was isolated from blood collected by orbital bleeding. The samples were immediately centrifuged, and the plasma was removed, diluted in water, and stored at −80 °C in a screw-top tube for later analysis. Methylmalonic acid was analyzed by gas chromatography–mass spectrometry with stable isotopic internal calibration to measure methylmalonic acid as previously described.[36,37] In vivo 1-13C-propionate oxidation was determined by collecting expired gas from mutant, control, and treated mice after the animals were injected by the intraperitoneal route with 200 micrograms of 1-13C-sodium propionate, using an adaptation of a method developed to study propionate oxidation in patients with methylmalonic and propionic acidemia.[38] The mice were placed into a respiratory chamber that contained a CO2 probe to allow the direct measurement of CO2 generated by each animal. An aliquot of expired air was removed from the chamber at each time point for analysis of 13C enrichment in CO2. The isotope ratio (13C/12C) of the expired gas was determined with a gas isotope ratio mass spectrometer (Metabolic Solutions, Nashua, NH). The percent dose metabolized at each time point was calculated as % dose metabolized = total 13C excreted (mmol/dose (mmol) × 100%).

In all instances, P values were considered significant if the value was <0.05. Differences in the survival between treated groups were analyzed using a χ2 test. The weights between treated and untreated mice, and differences in metabolite levels were assessed using a two-sided, two-tailed unpaired Student's t- test. The Kruskal–Wallis test was used to determine the statistical significance in measured propionate oxidation rates between groups.