AAT) After Portal Vein Injection of Recombinant Adeno-Associated Virus (Raav) Vectors
Total Page:16
File Type:pdf, Size:1020Kb
Gene Therapy (2001) 8, 1299–1306 2001 Nature Publishing Group All rights reserved 0969-7128/01 $15.00 www.nature.com/gt RESEARCH ARTICLE Stable therapeutic serum levels of human alpha-1 antitrypsin (AAT) after portal vein injection of recombinant adeno-associated virus (rAAV) vectors S Song1,4, J Embury2,4, PJ Laipis2,4, KI Berns1,4, JM Crawford2 and TR Flotte1,4 Departments of 1Pediatrics, 2Biochemistry and Molecular Biology, and 3Pathology, and 4Powell Gene Therapy Centre of the University of Florida Genetics Institute, University of Florida, Gainesville, FL, USA Previous work from our group showed that recombinant after portal vein injection of doses of 4 × 109 infectious units adeno-associated virus (rAAV) vectors mediated long-term (IU), a 10-fold lower dose than that required for similar levels secretion of therapeutic serum levels of human alpha-1 anti- of expression via the i.m. route. Serum levels greater than trypsin (hAAT) after a single injection in murine muscle. We 1 mg/ml were achieved at doses of 3 × 1010 IU. Southern hypothesized that hepatocyte transduction could be even blotting of liver DNA revealed the presence of circular episo- more efficient, since these cells represent the natural site of mal vector genomes. Immunostaining showed that trans- AAT production and secretion. To test this hypothesis, rAAV gene expression was scattered throughout the liver paren- vectors containing the hAAT cDNA driven by either the chyma. Similar results were obtained with a rAAV-CB-green human elongation factor 1 alpha promoter, the human cyto- fluorescent protein (GFP) vector. There was no evidence of megalovirus immediate–early promoter (CMV), or the CMV- hepatic toxicity. These data indicate that liver-directed rAAV- chicken beta actin hybrid (CB) promoter were injected into based gene therapy is effective in the murine model, and the portal or tail veins of adult C57Bl/6 mice. Potentially hence might be feasible for treatment of human AAT therapeutic serum levels of hAAT (600 g/ml) were achieved deficiency. Gene Therapy (2001) 8, 1299–1306. Keywords: hAAT; rAAV; gene therapy; hepatocyte Introduction ence with protein replacement also indicates a very low likelihood of eliciting immune responses to hAAT in Alpha 1-antitrypsin (AAT), a 52 kDa serine proteinase deficient patients.6 inhibitor, is normally secreted from hepatocytes and cir- Several vectors have been found to mediate detectable culates in the plasma, protecting lung elastin from degra- serum levels of hAAT in animal models.7–11 First-gener- dation by neutrophil elastase and related proteases. ation adenoviral vectors have been shown to give thera- Deficiency of AAT can lead to pan-acinar emphysema peutic levels of hAAT.12 Recently, high capacity aden- due to destruction of pulmonary interstitial elastin if ovirus,13,14 and rAAV vectors have demonstrated serum levels fall below 11 m (approximately feasibility of achieving stable levels greater than 800 1,2 800 g/ml). In a subset of patients homozygous for the g/ml. The study by our group indicated that a single PI*Z mutation, liver disease may also develop, appar- i.m. injection of a rAAV-CMV-AAT into murine muscle ently related to polymerization of the mutant protein could mediate stable expression within that range.15 within the endoplasmic reticulum of affected hepato- However, a relatively high dose of rAAV (4 × 1010 IU) 3–5 cytes. was required. Given the greater than 2000-fold size differ- AAT deficiency may be a good candidate for gene ther- ence between mice and adult humans, there is a signifi- apy because of the well-defined therapeutic endpoint and cant incentive to improve the efficiency of this process. wide therapeutic index. While levels of 800 g/ml seem Furthermore, the treatment of AAT deficiency-related necessary to avoid the development of clinically signifi- liver disease will likely be dependent upon transduction cant lung disease, much higher levels (ie 8 mg/ml or of a high percentage of hepatocytes. This is in contrast more) are observed during acute phase reactions without with gene therapy for lung disease, in which the percent obvious sequelae. Furthermore, earlier studies of protein transduction is not critical as long as the end result is replacement indicated that no adverse effects were seen a sufficiently high serum level. The current study was with serum hAAT levels greater than 100-fold above the designed to test the hypothesis that transduction of Ͼ therapeutic threshold (ie 80 mg/ml). Clinical experi- murine hepatocytes would be more efficient for expression and secretion of hAAT than skeletal myocytes. Portal vein injection of a rAAV-CB-hAAT vector resulted Correspondence: TR Flotte, University of Florida College of Medicine, Gene Therapy Center, 1600 SW Archer Road, Box 100266, Gainesville, in stable serum levels of hAAT at a 10-fold lower dose FL 32610–0266, USA than was required in the previous rAAV muscle Received 22 September 2000; accepted 21 December 2000 injection study. Delivery of rAAV-AAT to liver S Song et al 1300 Results hAAT expression in murine hepatocytes In order to determine the optimal promoter systems for in vivo studies, a series of rAAV-hAAT constructs utiliz- ing different expression cassettes were constructed (Figure 1). These included constructs with the human cytomegalovirus (CMV) immediate–early promoter (C- AT), the human elongation factor 1-␣ promoter (E-AT), or a hybrid CMV enhancer/chicken beta actin promoter/chimeric rabbit beta globin intron expression cassette (CB-AT). A version of C-AT with an intron and without neomycin-resistant gene was also generated (p43C-AT), as was a version of E-AT with a truncated version of the human elongation factor 1-␣ promoter (dE-AT). Figure 2 In vitro comparison of constructs. Mouse liver cells (HO15) Each construct was transfected into the SV40-T anti- were seeded and cultured in six-well plates overnight. Five micrograms of plasmid DNA from each construct were added to each well with trans- gen-transformed murine hepatocyte cell line, HO15, a fection reagent, LipofectAMINE. Two days after transfection, medium generous gift from Dr Janice Chou (National Institutes of was collected and analyzed for hAAT by ELISA. Control cells were trans- Health, Bethesda, MD, USA), and the expression of fected in the same way with a plasmid (UF5), which expresses GFP instead hAAT was assessed at 48 h using a sensitive and specific of hAAT and otherwise is identical to C-AT. Each bar represents the mean ELISA assay on cell medium supernatants. While the of six individual transfections. Constructs, C-AT, E-AT, and CB-AT were other promoters were active, the CB-AT construct described in Figure 1. Construct dE-AT is a derivative of E-AT from a deletion of a silencer region. Construct p43C-AT is similar to CB-AT showed a clear advantage in this cell line (Figure 2), except for using the CMV promoter instead of the CMV--actin hybrid expressing nearly two-fold greater levels than the C-AT promoter. contructs and five-fold greater levels than the E-AT con- structs. This could not be accounted for solely by the presence of an intron, as indicated by the comparison of the basic C-AT construct with an intron-containing version (p43CAT). Long-term in vivo expression after portal vein injection in young adult mice In order to determine the feasibility of achieving stable therapeutic levels (у800 g/ml) of hAAT in the serum, doses of each rAAV-hAAT vectors were injected into the portal veins or tail vein of cohorts of C57Bl/6 female mice. For the E-AT or C-AT vector, cohorts of mice were each injected with 4 × 109 infectious units (IU). For the CB-AT vector, two cohorts of mice were injected with either 3 × 1010 or 4 × 109 IU. Once again, the CB-AT vector was the most efficient, expressing sufficient hAAT to maintain serum levels of approximately 1 mg/ml in the high dose group (3 × 1010 IU per animal) for up to 52 Figure 3 Sustained therapeutic levels of hAAT secreted from mouse liver. weeks after a single injection (Figure 3). At the 4 × 109 Eight-week-old female C57 Bl/6J mice were injected through portal vein (PV) or tail vein (TV) with recombinant AAV vectors. Filled circle, CB- AT (PV), 3 × 1010 IU per animal (n = 2). Open circle, CB-AT (PV), 4 × 109 IU per animal (n = 4). Filled square, E-AT (PV), 4 × 109 IU per animal (n = 4). Open square, E-AT (TV), 4 × 109 IU per animal (n = 3). Open diamond, C-AT (PV), 4 × 109 IU per animal (n = 2). The serum levels of hAAT were monitored by ELISA at the time indicated. Means ± standard deviations are shown for most groups, except for those with n = 2. In the latter groups, means ± ranges are shown. IU dose, CB-AT still expressed hAAT at or near the thera- peutic level (600 g/ml) and was at least 50-fold more efficient than portal delivery of E-AT or C-AT. Interest- ingly, a cohort of mice treated with E-AT via tail vein injection had levels approximately five-fold lower than a Figure 1 Recombinant AAV-AAT vector cassettes. The C-AT (3.7 kb) cohort receiving an identical dose by direct portal vein and E-AT (4.4 kb) constructs are similar except that C-AT uses the CMV injection. This supports an earlier study in which a 10- ␣ promoter and E-AT uses the human elongation factor 1- (ELF) promoter. fold difference was observed.16 The CB-AT construct (3.8 kb) contains the CMV enhancer and chicken -actin promoter with a hybrid chicken-rabbit -globin intron.