Following the cloning of the cystic fibrosis (CF) gene, in vitro studies rapidly established the feasibility of gene therapy for this disease. Unlike ex vivo approaches that have been utilized for other genetic diseases such as adenosine deaminase deficiency, gene therapy for CF will likely require direct in vivo delivery of gene transfer vectors to the airways of patients with CF. Rence, major research efforts have been directed at the development of efficient gene transfer vectors that are safe for use in human subjects. Several vectors have now emerged from the laboratory for evaluation in clinical safety and efficacy trials in the United States and in the United Kingdom. Adenovirus-mediated gene transfer has been utilized for initial clinical safety and efficacy trials in the United States, while liposome-mediated gene transfer has been chosen for initial clinical safety and efficacy trials in the United Kingdom. The rationale and laboratory studies are reviewed leading to initial clinical safety and efficacy trials. Also reviewed are the currently available vectors for potential use in clinical studies, their advantages and disadvantages, and the promises and pitfalls of current gene therapy efforts f or CF in the United States focusing on adenovirus vectors in current clinical trials. (Chest 1995; 107.77S-83S)
AAV=adeno-associated virus; [CI.sup.-]=chloride; CF=cystic fibrosis; CFTR=cystic fibrosis transmembrane conductance regulator; E1=early region gene one; E3=early region gene three; MOI=multiplicity of infection
Key words: adeno-associated virus vectors; adeno virus vectors; airway epithelia; cationic liposomes; cystic fibrosis; gene therapy/transfer; molecular conjugates; retrovirus vectors
Cystic fibrosis (CF) is an autosomal recessive disorder affecting 1 in 2,500 live white births.[1] The disease is characterized by abnormal salt and water transport that leads to abnormal airway secretions, impaired mucociliary clearance, chronic bacterial infection, bronchiectasis, and premature death.[1,2] A variety of epithelial tissues are affected in this disease, including airway, pancreatic, sweat ductal, and gastrointestinal epithelia.[1,2] However, lung disease is the major cause of morbidity and mortality in this disorder. Accordingly, initial gene therapy efforts have been directed toward lung disease.
The CF gene, which codes for the CF transmembrane conductance regulator (CFTR) protein, was cloned in 1989,[3,4] and since that time, more than 300 mutations have been reported. The most common mutation is a three base pair deletion leading to deletion of phenylalanine (F) at position 508 ([delta]F508) of the gene. This mutation leads to abnormal intracellular processing and trafficking of the mutant CFTR, and ultimately to defective apical membrane chloride ([Cl.sup.-]) conductance in epithelial cells affected by this disorder.[5] Several investigators have proved that the protein product of CFTR is a cyclic adenosine monophosphate (cAMP)-mediated [Cl.sup.-] channel.[6-8]
Because CF is an autosomal recessive disorder, introduction of a normal copy of the gene into the host cell should result in restoration of normal [Cl.sup.-] transport function. In studies published by Drumm and colleagues[9] and Rich and colleagues,[10] introduction of a normal copy of the CFTR gene into CF epithelial cells using retrovirus and vaccinia virus vectors restored normal [Cl.sup.-] transport function to these cells. Subsequently, Olsen and colleagues" demonstrated persistence of normal [Cl.sup.-] transport function in CF epithelial cells for up to 6 months in culture following retrovirus-mediated gene transfer. These studies demonstrating restoration of normal [Cl.sup.-] transport following in vitro gene transfer of wild-type (ie, normal) CFTR have established the feasibility of gene therapy for CF.
Of the 46 human gene therapy protocols that had been reviewed by the Food Drug Administration by mid-July 1993, 38 of the protocols involved ex vivo approaches.[12] That is, cells are removed from an organ, placed in cell culture, genetically modified, usually with retrovirus vectors, enriched for the population of cells expressing the desired gene through the use of selectable markers, and then reinfused into the appropriate organ of the human body. This approach, although effective for disorders such as adenosine deaminase deficiency, will be virtually impossible to perform in CF in view of the dichotomously branching airways of the lung, which prohibit effective in vivo removal and reimplantation of airway epithelial cells. Therefore, gene therapy for CF will require direct in vivo delivery of vectors encoding wild-type CFTR to the airways of patients with CF. To be successful, this approach mandates the use of gene transfer vectors that are both efficient and safe for human use.
Current Gene Transfer Vectors
A variety of gene transfer vectors are potentially available for use in human clinical trials. These vectors can be broadly grouped into those that integrate or insert into the host cell genome (retroviruses and adeno-associated viruses [AAVs]) and those that function as extrachromosomal, or episomal, DNA in the nucleus leading to transient expression (molecular conjugates, liposomes, and adenoviruses). A list of vectors used in gene therapy research for CF, with their advantages and disadvantages, appears in Table 1, and they are described in more detail below.
Retrovirus Vectors
Retrovirus-mediated gene transfer is the method of gene transfer utilized in most currently approved human gene transfer trials.[13] While human immunodeficiency virus is the most commonly known retrovirus, most retroviruses are structurally similar. They are all RNA viruses that have two viral long terminal repeats that are important f or cellular integration and also contain promoter elements, a packaging signal, and a series of structural genes, gag, pol, and env. The gag gene encodes a precursor polypeptide that is cleaved to yield the capsid protein. The pol gene encodes reverse transcriptase and an enzyme involved in proviral integration. The env gene encodes the precursor to the envelope glycoprotein. A retrovirus vector is constructed by deleting the gag, pol, and env genes using restriction enzymes and inserting or cloning desired cDNAs (genes) into the remaining retrovirus genome.[14] Exogenous promoters may also be included in the sequences of the inserted gene or alternatively, the viral 5' long terminal repeat may be used to drive transcription of the inserted gene. Because these retrovirus vectors have had their viral genes deleted, they are replication-defective.
When a wild-type retrovirus infects a cell, glycoproteins on the virus particle bind to a specific cell surface receptor, which results in fusion and emptying of the nucleocapsid containing the RNA and its reverse transcriptase into the cytoplasm of the host cell. The RNA is reverse transcribed to form a cDNA, the provirus, which is translocated to the nucleus, where it integrates into the host cell chromosomes. The integrated provirus, through the normal process of DNA transcription, encodes new viral proteins and new viral RNA, which are assembled at the cell surface into new virus particles. When a replication-defective retrovirus vector infects a cell, its RNA is also reverse transcribed to form the provirus, which integrates into the host cell genome. However, unlike wild-type viruses, the integrated provirus from a retrovirus vector does not produce viral proteins or viral RNA. Instead, it encodes the inserted gene.
Because the retrovirus vector genome integrates into the host cell DNA, it offers a possibility for long-term expression, and perhaps a "cure." However, this process of random integration is associated with a risk of insertional mutagenesis and tumor formation. Also, the low rates of epithelial cell proliferation in normal human airways have made retrovirus vectors unattractive in in vivo gene transfer to the lung in view of their requirement for dividing cells.[15] Although a recent report suggests that airway epithelial cells in patients with CF may have significantly higher rates of cell turnover than measured in airways of normal subjects,[16] the experience with in vivo retrovirus-mediated gene transfer to the lung has been disappointing.
Adeno-Associated Viruses
Adeno-Associated virus is a naturally defective parvovirus that is nonpathogenic.[17,18] The two most common serotypes infecting humans are serotypes 2 and 3. These viruses require the presence of a helper virus in order to replicate or cause a lytic infection. When AAV infects a cell in the absence of a helper virus, it integrates into the host cell genome and becomes latent. Upon a subsequent wild-type adenovirus or wild-type herpes virus infection, the AAV genome can be rescued or excised from the chromosome to yield a lytic infection.
The AAV genome is small with a size of approximately 4.7 kb (kilobase pairs). Vector construction involves removal of all except approximately 0.2 kb of the virus genome.[18] The remaining 0.2 kb consists of the inverted terminal repeats at the 5' and 3' ends of the molecule. These inverted terminal repeats are important in the hairpin-type replication that is known to occur with this vector and also important for integration into the host cell genome. The region deleted includes the viral rep and cap genes, which are responsible for virus replication and nucleocapsid formation, respectively, and create approximately 4.5 kb of room for a cDNA insert. In the case of CFTR, which has a coding region of approximately 4.5 kb, this leaves little room for the use of either exogenous or endogenous promoters. Hence, AAV vectors encoding CFTR may actually exceed the size of the wild-type genome, leading to inefficient packaging and subsequent difficulties with virus production. Despite this limitation, in vitro correction of the CF [Cl.sup.-] permeability defect has been reported in CF airway epithelial cells using AAV vectors coding for wild-type and/or truncated CFTR cDNAs.[19,20]
A potential risk with the AAV vectors, given their ability to integrate into the host cell genome, is insertional mutagenesis. Unlike wild-type AAV, which integrates into specific sites on chromosome 19,[21] AAV vectors appear to integrate into multiple random sites.[22] Nevertheless, AAV-CFTR vectors are promising because they appear to infect and express in nondividing cells, perhaps existing temporarily in episomal form until DNA replication or repair occurs allowing integration.
DNA-Molecular Conjugates
When plasmid DNA is linked to polylysine and a receptor ligand, this forms a DNA-Molecular conjugate. This molecular conjugate may then bind to a specific cell surface receptor and enter the cell by receptor-mediated endocytosis. Addition of an inactivated adenovirus to this molecular complex using an antibody bridge results in more efficient gene transfer in vitro by virtue of the ability of the adenovirus to escape the endolysosomal pathway and increase DNA delivery to the nucleus.[2,3] Molecular conjugates have been shown to be efficient in airway epithelia in vitro. Moreover, recent data by Ferkol and colleagues[24,25] suggest that molecular conjugates that target the polymeric IgA receptor can deliver reporter genes to the airway epithelia in vitro and also in vivo following intravenous administration.
DNA-Molecular conjugates are attractive in view of the potential for targeting to specific tissues through the use of tissue-specific ligands. Moreover, these conjugates appear to be capable of infecting and expressing in quiescent cells. However, a major concern with DNA molecular conjugates is the large protein load that is delivered with the DNA. The immunologic considerations and effects of repetitive dosing with this transient expression vector have not been fully addressed.
Liposomes
Liposome-mediated gene transfer is a popular method by which to deliver DNA to cells. Perhaps even more common has been the use of liposomes to deliver encapsulated drugs to patients. The liposomes utilized in the pharmaceutical industry have largely been of the simple liposome or negatively charged variety. The use of negatively charged liposomes for gene transfer has been limited by a low efficiency of DNA encapsulation with subsequent low efficiency of gene transfer. The inclusion of fusion proteins and pH-sensitive lipids in the lipid bilayer of simple liposomes have improved gene transfer efficiency.[26] However, the use of cationic liposomes has brought liposome-mediated gene transfer to the forefront of gene therapy research.
Cationic liposomes are composed of cationic lipids such as N[1-(2,3-dioleoxy)propyl] N,N,N trimethylammonium (DOTMA), dimethyldioctadecyl-ammonium bromide (DDAB), or 3[beta][N-N',N'-dimethylamino ethane-carbamoyl] cholesterol (DC-cholesterol) mixed in varying molar ratios with the neutral phospholipid, dioleylphoshpatidylethanolamine (DOPE). In some instances, detergents and cholesterol may be added (as has been done with DC-cholesterol) to increase stability and possibly reduce in vitro toxicity. Cationic liposomes have generally been believed not to encapsulate DNA, although a recent publication suggests that encapsulation may occur at low DNA to liposome ratios.[27] Instead, negatively charged plasmid DNA binds to cationic liposomes to form a DNA-liposome complex. DNA-liposome complexes may enter cells through either fusion to the cell membrane or alternatively, via endocytosis. Current evidence suggests that endocytosis is the major route of uptake for DNA-liposome complexes.[26,28]
Liposome-mediated gene transfer has been shown to be efficient in vitro[26,28-31] and potentially in vivo[32-36] in a variety of cell types. It may also be efficacious in vivo in view of recent reports suggesting possible restoration of normal [Cl.sup.-] transport in transgenic mouse models of CF.[37,38] These reports, although controversial due to differences in electro-physiologic properties between murine and human airways, offer promising insights with regard to the potential role of liposome-mediated gene transfer for CF. Moreover, cationic liposomes have been associated with a low toxicity profile, making them very attractive as a gene transfer vector for human studies.[39-41] Currently, a clinical gene transfer safety and efficacy trial using liposome-mediated gene transfer is being performed in the nasal cavity of patients with CF in the United Kingdom. The issue of repetitive dosing with this transient expression vector system has not been addressed.
Adenovirus Vectors
Adenovirus virus vectors have anchored gene therapy efforts for CF in the United States. Adeno-viruses are icosahedral DNA viruses in which human serotypes 2 and 5 provide the backbone for current adenovirus vectors. Wild-type adenoviruses have a 36-kb genome, which for cloning purposes is conveniently divided into 100 map units of 360 base pairs each. The genome contains a series of early genes that are responsible for virus replication and antigen presentation and surveillance, and a series of late genes that encode viral structural proteins.[42-44] Current vectors based on adenovirus serotype 5 have generally had the early region one (E1) and early region three (E3) genes deleted.[45] Deletion of the El region makes the virus vector replication defective, while deletion of the E3 region creates enough room to insert the CFTR gene with a suitable promoter to drive transcription. Current adenovirus vectors encoding CFTR based on serotype 2,[46] which is approximately 90% homologous to scrotype 5, have only had the El region deleted such that vector genomes may actually exceed the size of the wild-type virus genome, leading to inefficient packaging or formation of viral particles and low infectious titers.
When an adenovirus vector infects a cell, it enters the cell by a process of receptor-mediated endocytosis.[47] Adenoviruses have an innate ability to escape the endolysosomal pathway, resulting in very efficient translocation of the DNA to the nucleus with subsequent high level episomal expression of exogenous DNA. Adenovirus vectors have been shown to be efficient in vitro and in vivo.[45,46,48,49] However, adenoviruses have known immunologic properties that may potentially limit their usefulness as human gene transfer vectors.[42-44] Moreover, because of transient expression, repetitive dosing will be required with this vector.
Current Approach to Gene Therapy in the United States
Current approaches to gene therapy in the United States are focused on the use of adenovirus-mediated gene transfer. Key issues to be addressed f or successful gene transfer studies in patients with CF include the following: (1) efficiency; (2) efficacy; and (3) safety.
Efficiency
A previous in vitro study by Johnson and colleagues[50] suggested that as few as 6 to 10% of CF airway epithelial cells. must be corrected in order to restore normal [Cl.sup.-] transport function to an entire epithelial sheet. Efficient gene transfer using adenovirus vectors has been reported in vivo in cotton rats.[49] However, in vivo adenovirus-mediated gene transfer to the airway epithelia of nonhuman primates tends to be of low efficiency and patchy.[51] Efficient gene transfer to human airway epithelia in vitro has also been reported.[46] As discussed below, published data from a clinical study in the nasal epithelium of patients with CF may offer some exciting insights into in vivo gene transfer of CFTR in patients with CF.[52]
Efficacy
The current working definition of efficacy is the physiologic demonstration of restoration of normal [Cl.sup.-] transport in CF epithelia following transduction of the wild-type human CFTR. In a previous study, Rosenfeld and colleagues[49] reported correction of the CF [Cl.sup.-] permeability defect in a CF pancreatic cell line in vitro and also in vivo CFTR protein expression in cotton rat airway epithelia. More recently, Rich and colleagues[46] have reported in vitro correction of the [Cl.sup.-] permeability defect in human CF airway epithelia with a multiplicity of infection (MOI) or number of infectious particles per cell as low as 0.1 to 1 using adenovirus-mediated transduction of CFTR.
To date, the mouse models for CF have not been used extensively in preclinical testing of adenovirus vectors. However, Grubb and colleagues (unpublished data, 1994) have demonstrated partial correction of the [Cl.sup.-] transport defect in the nasal epithelium of their CF mouse model[53,54] following in vivo adenovirus-mediated gene transfer to the nasal epithelium. Interestingly, a high MOI was required to partially restore [Cl.sup.-] transport. Moreover, no correction of sodium transport was observed. Recent evidence suggests that the results observed in the nasal studies of Grubb and colleagues may differ from observations in ongoing studies in patients with CF. In a series of three patients, Zabner and colleagues[52] have reported correction of [Cl.sup.-] transport f unction in the nasal epithelium of subjects enrolled in their clinical study of adenovirus-mediated gene transfer for CF at a MOI as low as 1.
Safety
Several safety concerns have been raised with adenovirus vectors, including contamination of virus stocks with recombinant replication competent virus, immune response, ectopic expression, and integration. Contamination of virus stocks can be minimized by screening for replication competent viruses. Perhaps a more worrisome concern is whether the presence of El genes in the affected epithelium of patients with CF, which has been reported to be as high as 10 to 15%,[55] Will lead to replication of adenovirus vectors in the host. For this reason, current human protocols evaluating safety and efficacy in patients have focused on the use of patients carrying antibodies to adenoviruses serotypes 2 or 5.
Wild-type adenoviruses are known to cause both an acute mixed cellular inflammatory response and a later lymphocyte predominant inflammatory response in rodents.[43,44] The acute response may be cytokine-mediated, while the late response is probably mediated by cytotoxic T cells.[43,44] Similarly, Engelhardt et al[51] and Trapnell et al[56] have been able to demonstrate both an acute neutrophil response and a late lymphocyte predominant inflammatory response in cotton rat and the baboon lung following direct delivery of adenovirus serotype 5-based vectors to the airways. Interestingly, the inflammatory responses were seen only when high doses of adenovirus vector were delivered to the alveolar region of the lung. In unpublished studies, we have seen similar dose-related inflammatory responses following adenovirus-mediated gene transfer to murine lungs (D. Thomas et al, unpublished results, 1994). These results suggest that the therapeutic to toxic ratio with adenovirus vectors may be small.
The data with regard to toxicity following ectopic expression of CFTR are mixed. Adenovirus-mediated gene transfer may lead to high-level expression in many different cell types in the airways. In a recent study by Stutts and colleagues,[57] clonal fibroblast cell lines overexpressing CFTR grew more slowly and had different electrophysiologic properties, ie, depolarized membrane potentials, when compared with fibroblasts that expressed CFTR at low levels or not at all. In contrast, transgenic mice that overexpress CFTR do not differ from control mice that do not overexpress the gene in their lung development, somatic growth characteristics, or reproductive function.[58] Hence, overexpression of CFTR may be manifested primarily in cells of nonepithelial origin that lack distinct apical and baso-lateral domains.
Adenovirus serotypes 2 and 5 have been used to efficiently transform human and rodent cell lines at frequencies up to 100-fold greater than obtained using calcium phosphate.[48] In general, transformation has been mediated through the E1 region genes, although substitution of simian virus 40 (SV40) sequences for the El genes will also enable stable transformation. The reports of E1 sequences present in the airway epithelial cellular genomes of CF and normal individuals[55] suggest that stable integration may also occur in airway epithelia. In preliminary in vitro studies from our laboratory, integration may occur in airway cells following adenovirus-mediated gene transfer at frequencies of up to 5 in 10,000 (J. Olsen et al, unpublished data, 1994). Given the requirement for repetitive dosing with in vivo adenovirus-mediated gene transfer of CFTR, the risk of insertional mutagenesis may be considerable.
Human safety and efficacy studies are currently underway at five CF centers. These studies are using low doses of adenovirus vectors in small numbers of patients with CF. If no toxicity is seen at the lower doses, the doses will be increased at subsequent stages of these trials. The initial results reported from the trial at the University of Iowa appear to be promising. One severe adverse reaction, that may have been cytokine-mediated, has been reported in one study involving the lower airways.
Conclusions
Major research efforts have been conducted to develop suitable gene transfer vectors for CF. Although many of these vector systems are still under development and are in need of improvement, initial clinical safety and efficacy trials for gene therapy of CF have begun. Liposome-mediated gene transfer has been the vector system selected for initial clinical studies in the United Kingdom, while clinical safety and efficacy trials in the United States all utilize adenovirus-mediated gene transfer. In this article, we have shown that adenovirus-mediated gene transfer to the airway epithelium can be efficient in vivo, although it may be rather inhomogeneous. Moreover, this method of gene transfer may be efficacious for CF based on evidence from mouse models and ongoing in vivo studies in humans. However, lung inflammation is a major concern, with the use of high doses of adenovirus vectors. Moreover, the therapeutic to toxic ratio with repetitive dosing has not been defined. if gene therapy for CF is to become a reality, then continued development of more efficient vectors with improved safety profiles will be required.
(*) From the Division of Pulmonary and Critical Care Medicine, Department of Medicine, and the Cystic Fibrosis/Pulmonary Research and Treatment Center, The Universiiy of North Carolina at Chapel Hill. Reprint requests: Dr. Johnson, Div of Pulmon/CCM, CB 7020, Rm 724, Burnett Womack Bldg, UNC-CH, Chapel Hill, NC 27599-7020
COPYRIGHT 1995 American College of Chest Physicians
COPYRIGHT 2004 Gale Group
Contact
Home
A
Aagenaes syndrome