Gene-Based Therapy for Alpha-1 Antitrypsin Deficiency

Abstract

Alpha-1 antitrypsin Deficiency (AATD) has been an attractive target for the development of gene therapy because it is a common single gene disorder, for which there would appear to be significant benefit to be gained for lung disease patients by augmentation of plasma levels of wild-type (M) alpha-1 antitrypsin (AAT). While a significant proportion of patients also have liver disease, which is unlikely to be benefitted by augmentation, the potential to treat or prevent lung disease by replacement of plasma levels to at least 11 microMolar (571mcg/ml) is the basis upon which several protein replacement therapies have been licensed for human use. Further enhancing the likelihood of success of gene therapy is the fact that the AAT coding sequence is relatively short and the protein appears to function primarily in the plasma and extracellular space. This means that AAT production from any cell or tissue capable of secreting it could be useful therapeutically for augmentation. Based on these considerations, attempts have been made to develop AAT therapies using nonviral gene transfer, gammaretrovirus, recombinant adenovirus (rAd), and recombinant adeno-associated virus (rAAV) vectors. These have resulted in three phase I clinical trials (one of cationic liposome, one of rAAV2, and one of rAAV1) and one phase II clinical trial (with rAAV1). The results of the latter trial, while promising, demonstrated levels were only 3 to 5% of the target range. This indicates the need to further increase the dose of the vector and/or to increase the levels to within the therapeutic range.

Keywords :

• gene therapy
• alpha 1 antitrypsin
• AAV
• AAT
• antitrypsin
• gene Therapy
• adeno-associated viral vector

Introduction

Alpha-1 antitrypsin Deficiency (AATD) has been an attractive target for the development of gene therapy because it is a common single gene disorder, for which there would appear to be significant benefit to be gained for lung disease patients by augmentation of plasma levels of wild-type (M) alpha-1 antitrypsin (AAT). While a significant proportion of patients also have liver disease, which is unlikely to be benefitted by augmentation, the potential to treat or prevent lung disease by augmentation of plasma levels to at least 11 microMolar (571 mcg/ml) is the basis upon which several protein augmentation therapies have been licensed for human use.

Further enhancing the likelihood of success of gene therapy is the fact that the AAT coding sequence is relatively short and the protein appears to function primarily in the plasma and extracellular space. This means that AAT production from any cell or tissue capable of secreting it could be useful therapeutically for augmentation. Based on these considerations, attempts have been made to develop AAT therapies using nonviral gene transfer, gammaretrovirus, recombinant adenovirus (rAd), and recombinant adeno-associated virus (rAAV) vectors. These have resulted in three phase I clinical trials (one of cationic liposome, one of rAAV2, and one of rAAV1) and one phase II clinical trial (with rAAV1). The results of the latter trial, while promising, demonstrated levels were only 3 to 5% of the target range. This indicates the need to further increase the dose of the vector and/or to increase the levels to within the therapeutic range.

General considerations

Substantial research on AAT gene transfer has been performed with nearly every vector type. This is most likely because AAT is a secreted protein whose serum levels can be readily assayed. This allows for direct quantitative comparisons of the level and duration of gene transfer with various doses, routes, and vector subtypes, much as one could obtain from a reporter gene. This property also allows for the definition of a critical endpoint for therapeutic development, namely the augmentation of plasma levels to 11 microMolar or 571 micrograms/ml, which has been considered therapeutic for the prevention of progression of AATD-related lung disease. Although there has been substantial debate about this endpoint, it is supported by both natural history studies of compound heterozygote patients (1) and by retrospective analyses of patient mortality while on protein augmentation therapy. Finally, since AAT appears to be readily secreted from a number of different organs, allowing augmentation of plasma levels to be potentially feasible after gene delivery to liver, muscle, lung, or pleura.

The endpoint for AATD-related liver disease is much less certain, reflecting the even greater uncertainty about liver disease pathogenesis. In general, there is consensus that the presence of mutant protein in hepatocytes, either due to its misfolded state, to its retention within hepatocytes, to the polymerized state of most common AAT mutant (PiZ) protein, or to some combination of these factors. In any of these instances, the molecular objective for treatment or prevention of AATD liver disease is knockdown of the mutant protein. The degree of knockdown required for therapeutic effect is also uncertain, but the observation that liver disease is not evident in MZ (heterozygous wild-type/Z mutant) patients has led to the conjecture that 50% knockdown could be clinically beneficial.

A further level of complexity is added in simultaneous gene therapy for both AATD lung and liver disease. Such an approach requires either a combination of a knockdown vector with augmentation from a different organ, or a selective, allele-specific system in which the knockdown vector is active against the endogenous transcript but not against the therapeutic augmentation allele. The latter approach will be described next.

Nonviral systems

The earliest clinical trial of AAT gene therapy was performed using cationic liposomes to deliver the M-AAT cDNA to the nasal epithelium of patients with AAT deficiency (2). This study, published in 2000 was preceded by substantial preclinical evaluation (3,4). In the clinical trial itself, the M-AAT gene was delivered in a plasmid-cationic liposome complex to one nostril of each of five subjects with AAT deficiency, with the other nostril serving as a control. AAT protein concentration in nasal lavage fluid increased in the transfected nostril, but not in the control, peaking at 5 days after instillation, and IL-8 concentration in the nasal lavage were decreased, indicating biological activity (Figure 1). It was unclear from this early study how this could be translated into a therapy for the lung disease, particularly when another study showed that DNA-cationic liposome delivery to the lung by aerosol resulted in a dose-limiting flu-like toxicity syndrome (5).

Figure 1.  Nasal lavage fluid AAT and IL-8 concentrations after gene transfer of AAT with a cationic liposome vector. Concentrations of AAT normalized to total protein concentration in nasal lavage fluid from patients while receiving no therapy, while receiving weekly intravenous AAT protein therapy, and at the time of peak AAT concentration after transfection with the AAT gene. Means and SD for 10 normal subjects are shown for comparison. Among patients not receiving therapy, AAT concentrations were low, as expected (p, < 0.05 versus normal, unpaired t test). AAT protein therapy increased AAT levels to normal. Transfection with the AAT gene increased AAT levels to about one-third of the normal mean. Bottom: Concentrations of IL-8 normalized to total protein concentration in the same nasal lavage fluid samples for which AAT data are shown (top). Means and SD for 10 normal subjects are shown for comparison. Among patients not receiving therapy, IL-8 concentrations were significantly higher than normal (p, 0.05, unpaired t test). Although receiving AAT protein therapy, mean IL-8 levels were higher than while not receiving therapy, but the difference was not significant. In contrast, transfection with the AAT gene decreased IL-8 concentrations to normal (p, 0.05 versus no therapy, Wilcoxon paired test). [With permission from Brigham, et al. Hum Gene Ther 2000 May 1;11(7):1023–1032. PMID:10811231].

Recently, there has been an effort to utilize synthetic siRNA delivery as a therapy for AAT liver disease (personal communication). Although these studies are in early stages, there have been indications that knockdown of the Z-AAT protein expression is feasible, and this could progress to clinical translation.

Retrovirus-mediated gene transfer

Much of the early work in AAT gene therapy was based on the concept that one might explants patient hepatocytes, expand and transduce them in culture with a gammaretrovirus vector, and then reinfuse or reimplant them as autologous tissue back into the patient from whom they were obtained. Early utilization of this platform in patients was achieved with LDL-receptor deficiency (6). The work with the human AAT gene was undertaken and likewise demonstrated feasibility in animal models (7) (Figure 2). The invasive nature of this process, and the subsequent demonstration of tumorigenicity of gammaretrovirus vectors (8,9) have limited the further clinical development of this approach. Similar concepts have been undertaken with lentivirus vectors, but similarly none have been attempted in human trials.

Figure 2.  Serum hAAT levels of transplantation of retrovirus tranduced hepatocytes in two individual canines (A1 and A2). In vivo hAAT production in dogs Al and A2 after transplantation of transduced hepatocytes. The serum concentrations of hAAT were determined before and after transplantation of 3.8 × 108 and 6.4 × 101 in animals Al (solid circles) and A2 (open triangles), respectively. Each sample was analyzed in duplicate. [Reproduced with permission from Kay MA, et al. Proc Natl Acad Sci USA 1992 Jan 1;89(1):89–93. PMID:1729724].

Recombinant adenovirus (rAd)

Recombinant adenovirus (rAD) vectors enjoyed popularity for high-level transient expressions of transgenes, and were thought to have a particular advantage when delivering genes to the lung. rAd-AAT vectors were developed and used in preclinical trials, but limitations of this system in the clinical setting, primarily due to a robust innate immune response, have dampened enthusiasm for further pursuing this system.

Recombinant adeno-associated virus (rAAV)

Recombinant adeno-associated virus (rAAV) vectors, were first tested in humans in 1995 in a trial in adult CF patients with mild lung disease (10). Efforts to apply rAAV to gene therapy for AAT deficiency began in the late 1990s with a variety of studies testing direct hepatic delivery or intramuscular delivery of the gene for ectopic secretion of the wild-type M-AAT protein (11,12). The ability to transfer genes for secreted proteins to muscle and have muscle efficiently secrete those proteins had been demonstrated a few years earlier with erythropoietin (13) and factor IX (14,15).

In the study by Song et al. (12) a rAAV serotype 2 AAT vector mediated stable expression of M-AAT in C57Bl6 mice for many months without decrement. These studies were possible because C57Bl6 mice are tolerant for human AAT, and yet the murine alpha-1 proteinase inhibitor protein does not cross-react with polyclonal human anti-AAT antibodies used for the ELISA assay. A subsequent study showed persistence of M-AAT a full year after a single IM injection. Levels were slightly higher in severe combined immune deficiency (SCID) mice. This group was initially included because of the potential that an immune response was limiting. However, subsequent studies showed that the SCID defect in these mice was due to deficiency in a nuclear enzyme called DNA-dependent protein kinase catalytic subunit (DNA-PKcs).

DNA-PKcs deficiency results in immune deficiency because it is needed for the type of recombinant required for creating V-D-J junctions in immunoglobulins and T-cell receptors. However, that same mechanism (non-homologous end-joining) appears to be involved in the formation of the stable rAAV episomal concatemers (16,17). In the presence of DNA-PKcs, the normal processing occurs, but in its absence vector integration is increased. This is probably of little relevance to the utility of rAAV-AAT in muscle, but it may have implications for understanding the processing of rAAV genomes in the cell in general, and indicates that the relative non-carcinogenicity of rAAV is likely due to a very specific processing mechanism within the cell.

Given the promising preclinical data with rAAV2-AAT, formal preclinical pharmacology, toxicology, and biodistribution studies were performed and a phase I trial was initiated in AAT deficient adults. Ultimately, this study showed only transient low-level expression, but an excellent safety profile was documented. Subsequent data indicated that other serotypes, most notably rAAV1, were capable of much higher levels of secretion. A phase I trial of rAAV1-AAT was then initiated and showed a much higher and more sustained level of expression, even in the face of an effector T-cell response against the capsid.

By this point, a transition of production of the clinical-grade vector was made to a herpes simplex virus (HSV)-based system, and the dose was increased as a phase 2 study was initiated. In the phase 2 study, M-AAT levels peaked at 30 days, and then coincident with a spike in muscle enzymes, decreased by about half before reaching a steady plateau at 3 to 5% of the therapeutic target (Figure 3). Based on this finding a second phase 2 trial is contemplated in which the dose and potency might be further enhanced using an isolated limb infusion method and a dose escalation.

Figure 3.  Serum M-specific α₁-antitrypsin (AAT) concentration after injection of rAAV1-CB-hAAT produced by plasmid transfection (TFX) or the herpes simplex virus (HSV) method. Values shown represent means±SD. The dose of vector administered to subjects is indicated in the figure legend. Values for the TFX group are from a previous study. Values for the 6×1011 VG/kg HSV group do not include results for subject 303, who had an AAT phenotype of SZ; the monoclonal antibody used to determine serum M-specific AAT concentrations has little cross-reactivity with Z-type AAT but cross-reacts strongly with S-type AAT, causing results for this assay in this subject to be spuriously high [Reproduced with permission from Flotte et al. Hum Gene Ther 2011 Oct;22(10):1239–1247. Epub 2011 Aug 24. PMID:21609134].

More recently, alternative modes of delivery have been investigated (18), most notably intrapleural (IP) delivery or rAAV5 and rAAVrh10 vectors (19-21). The latter mode of delivery appears to be the most potent studied to date in the mouse (Figure 4). Although the route of delivery is somewhat more invasive than direct IM or IV delivery, it is feasible and further investigations are ongoing.

Figure 4.  Pharmacokinetics and biological activity of rAAVrh10AAT after intrapleural delivery in C57Bl6 mice. Time course and function of serum human a1AT levels in mice following intrapleural administration of AAVrh.10ha1AT vector. (A) Timecourse. The AAVrh.10 human A1AT vector (1011 genome copies) was administered intrapleurally to male C57BL/6 mice (n = 4/group). Serum human A1AT levels were measured at the time of vector administration (day 0) and at days 4 to 168 following vector administration. Serum A1AT levels were measured by ELISA; values shown are means F standard error. (B) Function. Assessment was made of inhibition of neutrophil elastase by human a1AT in the serum following intrapleural administration of AAVrh.10 human A1AT 6 weeks previously. The% inhibition of elastase activity is shown on the ordinate. The top abscissa shows the serum dilution, and the bottom abscissa shows the amount of human A1AT (in ug) present in the positive control, i.e., spiked naive serum, or present in the serum of mice injected with the AAVrh.10 human A1AT vector. Undiluted mouse serum following AAVrh.10 A1AT administration had human a1AT levels of 2 ug/ul. The serum of naive mice, which does not contain human A1AT, was used as a negative control; as expected, the endogenous murine A1AT was capable of inhibiting neutrophil elastase. [Reproduced with permission from De BP, et al. Mol Ther 2006 Jan;13(1):67–76. Epub 2005 Nov 2. PMID: 16260185].

rAAV for combined liver and lung gene therapy

Early attempts at vector-mediated knockdown of mutant AAT, involved the use of ribozymes, or short-hairpin RNAs (shRNA), the former being demonstrated with SV40-based vector and the latter with a rAAV8 vector (22,23). Concerns arose about the safety of shRNA over-expression in the liver (24).This pointed to the potential utility of synthetic miRNA-based vectors.

Recently, two laboratories have demonstrated rAAV-based delivery of synthetic miRNAs have the ability to down-regulate the endogenous mutant human AAT gene in a commonly used transgenic mouse strain that expresses hAAT from its own promoter (25,26). In each case, a version of the wild-type gene was expressed simultaneously. Mueller et al. devised a vector in which the placement and copy number of the miRNA templates were optimized and expressed within the same transcript as a version of the M-AAT coding sequence, which had been rendered resistant to the synthetic miRNA by codon usage changes which did not change the amino acid sequence of the protein.

Utilizing that strategy, systemic IV delivery of the construct in a rAAV9 capsid resulted in an average of 80% knockdown of Z-AAT with simultaneous augmentation of therapeutic levels of M-AAT. Both these therapeutic effects were maintained at stable levels for over 3 months (Figure 5). The safety concerns were addressed by performing miRNA profiles in which the levels of all endogenous miRNAs were also assessed. No perturbations of the endogenous miRNA profile were observed in the vector-treated animals.

Figure 5.  Simultaneous allele-specific knockdown of PiZ AAT and augmentation of PiM AAT in a transgenic mouse model expressing human Z-AAT. Transgenic mice expressing the human PiZ allele were injected with 1 × 1012 virus particles or rAAV9 expressing miRNAs against AAT and a de-targeted cMyc tagged wild-type M-AAT cDNA under the control of the hybrid chicken β-actin promoter via the tail vein. (a) Serums from each cohort were collected on a weekly basis and were used to assess Z-AAT concentration by ELISA. Serum Z-AAT levels at each timepoint are expressed as a percent knockdown as compared to the rAAV9-GFP cohort by using a Z-specific AAT ELISA and M-AAT levels are calculated by using an ELISA to quantify the cMYC tag on the wild-type protein. Data are expressed as group means ±SEM (n = 6). Statistical significance was set at * p ≤ 0.05 as determined by a two-way ANOVA comparing each treatment group to the control rAAV-GFP group. Blue dashed line in the upper panel indicates therapeutic levels of wild-type PiM AAT as determined by the FDA and therapeutic knockdown of PiZ protein in the lower panel as determined by achieving levels expected in a PiZ heterozygous status. Total RNA from mouse livers was used to assay for the presence of the either. Data are expressed as group means ±SEM (n = 6).* p ≤ 0.05 as determined by a two-way unpaired Student's t-test. AAT, α-1 antitrypsin; ANOVA, analysis of variance; cDNA, complementary DNA; ELISA, enzyme-linked immunosorbent assay; CB-GFP, chicken β-actin–green fluorescent protein; FDA, Food and Drug Administration; mRNA, messenger RNA; miRNA, microRNA; ND, not detected; qRT-PCR, quantitative reverse transcriptase-PCR; rAAV, recombinant adeno-associated virus. [Adapted with permission from Mueller, et al. Mol Ther. 2012 Mar;20(3):590-600. doi: 10.1038/mt.2011.292. Epub 2012 Jan 17. PMID:22252449]

Summary

In summary, gene therapy for AAT deficiency has been developed using multiple different vector systems, including cationic liposomes, retroviruses, rAd, and several different rAAV serotypes. Vectors have been administered by several different routes of delivery, as well. Most recently, vectors capable of allele-specific down-regulation of Z-AAT with simultaneous robust augmentation of M-AAT have been evaluated in vivo and found to be efficacious, stable, and safe in a mouse model.

The current active clinical development program for augmentation of M-AAT involves IM administration of a rAAV vector, a strategy for serum protein augmentation that has been demonstrated to be effective for a number of other secreted proteins, most notably for secretion of lipoprotein lipase (LPL) in clinical trials supporting the licensure of the first gene therapy product to be recommended by the European Medicines Agency (EMA) for approval in Europe. These facts would appear to bode well for the future of rAAV-based human gene therapy.

Declaration of Interest Statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.