Clinical Review: Gene-based therapies for ALI/ARDS: where are we now?

Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) confer substantial morbidity and mortality, and have no specific therapy. The accessibility of the distal lung epithelium via the airway route, and the relatively transient nature of ALI/ARDS, suggest that the disease may be amenable to gene-based therapies. Ongoing advances in our understanding of the pathophysiology of ALI/ARDS have revealed multiple therapeutic targets for gene-based approaches. Strategies to enhance or restore lung epithelial and/or endothelial cell function, to strengthen lung defense mechanisms against injury, to speed clearance of infection and to enhance the repair process following ALI/ARDS have all demonstrated promise in preclinical models. Despite three decades of gene therapy research, however, the clinical potential for gene-based approaches to lung diseases including ALI/ARDS remains to be realized. Multiple barriers to effective pulmonary gene therapy exist, including the pulmonary architecture, pulmonary defense mechanisms against inhaled particles, the immunogenicity of viral vectors and the poor transfection efficiency of nonviral delivery methods. Deficits remain in our knowledge regarding the optimal molecular targets for gene-based approaches. Encouragingly, recent progress in overcoming these barriers offers hope for the successful translation of gene-based approaches for ALI/ARDS to the clinical setting.


Gene therapy: opportunities in ALI/ARDS
Gene-based therapy involves the insertion of genes or smaller nucleic acid sequences into cells and tissues to replace the function of a defective gene, or to alter the production of a specifi c gene product, in order to treat a disease. Gene therapy can be classifi ed into germline and somatic gene therapies. Germline approaches modify the sperm or egg prior to fertilization and confer a stable heritable genetic modifi cation. Somatic gene approaches use gene therapy to alter the function of mature cells. Commonly used somatic gene therapy strategies include the overexpression of an existing gene and/or the insertion of smaller nucleic acid sequences into cells to alter the production of an existing gene.
ALI/ARDS may be suitable for gene-based therapies as it is an acute but relatively transient process [8], requiring short-lived gene expression, obviating the need for repeated therapies and reducing the risk of an adverse immunological response. Th e distal lung epithelium is selectively accessible via the tracheal route of administration, allowing targeting of the pulmonary epithelium [9]. Th e pulmonary vasculature is also relatively accessible, as the entire cardiac output must transit this circulation. Antibodies that bind antigens selectively expressed on the pulmonary endothelial surface can be complexed to gene vectors to facilitate selective targeting following intravenous administration [10]. It is also possible to use gene-based strategies to target other cells central to the pathogenesis of ALI/ARDS, such as leuko cytes and Abstract Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) confer substantial morbidity and mortality, and have no specifi c therapy. The accessibility of the distal lung epithelium via the airway route, and the relatively transient nature of ALI/ ARDS, suggest that the disease may be amenable to gene-based therapies. Ongoing advances in our understanding of the pathophysiology of ALI/ARDS have revealed multiple therapeutic targets for genebased approaches. Strategies to enhance or restore lung epithelial and/or endothelial cell function, to strengthen lung defense mechanisms against injury, to speed clearance of infection and to enhance the repair process following ALI/ARDS have all demonstrated promise in preclinical models. Despite three decades of gene therapy research, however, the clinical potential for gene-based approaches to lung diseases including ALI/ ARDS remains to be realized. Multiple barriers to eff ective pulmonary gene therapy exist, including the pulmonary architecture, pulmonary defense mechanisms against inhaled particles, the immunogenicity of viral vectors and the poor transfection effi ciency of nonviral delivery methods. Defi cits remain in our knowledge regarding the optimal molecular targets for genebased approaches. Encouragingly, recent progress in overcoming these barriers off ers hope for the successful translation of gene-based approaches for ALI/ARDS to the clinical setting. fi bro blasts [11]. Furthermore, gene-therapy-based approaches off er the potential to selectively target diff erent phases of the injury and repair process. Th e potential to target specifi c aspects of the injury and repair processes such as epithelial-mesenchymal transition, fi brosis, fi brinolysis, coagulopathy and oxidative stress with these approaches is also clear.

Current gene-based approaches
Gene therapy requires the delivery of genes or smaller nucleic acid sequences into the cell nucleus using a carrier or vector. Th e vector enables the gene to overcome barriers to entry into the cell, and to make its way to the nucleus to be transcribed and translated itself or to modulate transcription and/or translation of other genes. Both viral and nonviral vector systems have been developed (Table 1).

Viral vector-delivered gene therapy
Viral vectors are the most eff ective and effi cient way of getting larger nucleic acid sequences, particularly genes, into cells (Table 1). Th e viral genome is modifi ed to remove the parts necessary for viral replication. Th is segment is then replaced with the gene of interesttermed a transgene -coupled to a promoter that drives its expression. Th e modifi ed genome is then encapsulated with viral proteins. Following delivery to the target site, the virus binds to the host cell, enters the cytoplasm and releases its payload into the nucleus (Figure 1). Th e size of trans gene that can be used depends on the capsid size. A number of diff erent viral vectors have been used in preclinical lung injury studies to date.

Adenoviral vectors
Adenoviruses have double-stranded DNA genomes, have demonstrated promise in preclinical models [12,13] and are well tolerated at low to intermediate doses in humans [14,15]. Advantages include their ease of production, the high effi ciency at which they can infect the pulmonary epithelium [14,16] and that they can deliver relatively large transgenes. A disadvantage of adenoviruses is their immunogenicity, particularly in repeated doses [14]. Newer adenoviral vectors, in which much of the immuno genicity has been removed, hold promise [17]. While adenovirus-mediated gene transfer in the absence of epithelial damage is relatively ineffi cient [18], this may be less of a problem in ALI/ARDS that is characterized by widespread epithelial damage.

Adeno-associated virus vectors
Adeno-associated viruses (AAVs) are single-stranded DNA parvoviruses that are replication defi cient [19]. A substantial proportion of the human population has been exposed to AAVs but the clinical eff ects are unknown.
AAV vectors have a good safety profi le, and are less immunogenic compared with other viruses, although anti bodies do develop against AAV capsid proteins that can compromise repeat administration. AAV vectors can insert genes at a specifi c site on chromosome 19. Th e packaging capacity of the virus is limited to 4.7 kb, restricting the size of the transgene that can be used. AAVs are less effi cient in transducing cells than adenoviral vectors. Successful AAV vector gene transfer has been demon strated in multiple lung cell types including lung progenitor cells, in both normal and naphthaleneinduced ALI lungs [20]. AAV serotypes have specifi c tissue tropisms, due to diff erent capsid proteins that bind to specifi c cell membrane receptors. AAV-5 [21] and AVV-6 [22] exhibit enhanced tropism for the pulmonary epi thelium [21,22]. AAVs can transduce nondividing cells and result in long-lived transgene expression. AAV vectors have been used in clinical trials in cystic fi brosis patients, underlining their safety profi le [23,24].

Lentivirus vectors
Th ese RNA viruses can transfect nondividing cells such as mature airway epithelial cells [25]. Th e virus stably but randomly integrates into the genome and expression is likely to last for the lifetime of the cell (~100 days). Th e transgene can be transmitted post mitosis, and there is also a risk of tumorigenesis if the transgene integrates near an oncogene. Th e development of leukemias in children following gene therapy for severe combined immunodefi ciency highlights this risk [26,27]. While lentiviral vectors may be useful to correct a gene defi ciency associated with increased risk of ALI, the long-lived gene expression of lentiviral delivered genes may be more suitable for chronic diseases than for ALI/ARDS.

Nonviral gene-based strategies
Nonviral delivery systems, while generally less effi cient than viral vectors in transfecting the lung epithelium, are increasingly used to deliver smaller DNA/RNA molecules (Table 1). Strategies include the use of DNA-lipid and DNA-polymer complexes and naked DNA/RNA oligonucleotides, such as siRNA [28], decoy oligo nucleo tides [29] and plasmid DNA [30]. Nonviral delivery systems are less immunogenic than viral vector-based approaches, and can be generated in large amounts at relatively low cost.

Plasmid transfer
Plasmid vectors are composed of closed circles of doublestranded DNA. As naked and plasmid DNA contain no proteins for attachment to cellular receptors, there is no specifi c targeting to diff erent cell types and thus it is essential that the DNA is placed in close contact with the desired cell type. Th ese limitations make this approach less relevant clinically.

Nonviral DNA complexes
Th e therapeutic DNA is held within a sphere of lipids, termed a lipoplex, or within a sphere of polymers, such as polyethyleneimine, termed a polyplex. Lipoplexes and polyplexes act to protect the DNA, facilitate binding to the target cell membrane and also trigger endocytosis of the complex into the cell, thereby enhancing gene expression. Th ese systems can be modifi ed to include a targeting peptide for a specifi c cell type, such as airway epithelial cells [31]. Th ese complexes effi ciently and safely transfect airway epithelial cells [31], and they have demonstrated promise in human studies [32].

DNA and RNA oligonucleotides
siRNAs are dsRNA molecules of 20 to 25 nucleotides that can regulate the expression of specifi c genes. Specifi c siRNAs reduce infl ammation-associated lung injury in Table 1.

Approach Advantages Disadvantages Examples
Viral vector-delivered gene therapy

Adenoviral vectors
Relatively easily produced Immunogenic [14] Adenoviral transfer of genes for a surfactant (dsDNA genome) Effi ciently transfect lung enzyme [49], angiopoietin-1 [51], HSP-70 [52], epithelium [14,16] apolipoprotein A-1 [53], and Na + ,K + -ATPase pump Can deliver larger genes [55] genes attenuate experimental ALI Well tolerated in lower doses [1,3] Adenoviral delivery of IL-10 gene attenuates zymosan ALI at low doses, but is harmful at high doses [58] Adeno-associated virus Good safety profi le; less Limited transgene size AAV vector gene transfer demonstrated in multiple vectors (ssDNA genome) immunogenic Diffi cult to produce in large lung cell types including progenitor cells in both Inherently replication defi cient quantities normal lungs and following naphthalene-induced AAV-5 and AAV-6 lung epithelial ALI [20] tropism [10,11] Transduce nondividing cells Long-lived gene expression Used in clinical trials for CF [12,13] Lentivirus vectors Transduce nondividing cells [25] Oncogenesis risk due to Lentiviral transfer of shRNA to silence CD36 gene (RNA genome) Integrate stably but randomly integration into genome expression suppresses silica-induced lung fi brosis into the genome [26,27] in the rat [35] Nonviral gene-based strategies Plasmid transfer (closed Easily produced at low cost No specifi c cell targeting Electroporation-mediated gene transfer of the dsDNA circles) Very ineffi cient Na + ,K + -ATPase rescues endotoxin-induced lung injury [60] Nonviral DNA complexes Complexes protect DNA Less effi cient than viral vectors Cationic lipid-mediated transfer of the Na + ,K + -(lipoplexes or polyplexes) Complexes facilitate cellular ATPase gene ameliorated high-permeability targeting [31] pulmonary edema [59] Lipoplex-delivered IL-10 gene decreased CLP-induced ALI [61] Systemic cationic polyethylenimine polyplexes incorporating indoleamine-2,3-dioxygenase decreased ischemia-reperfusion ALI [62] DNA and RNA Easily produced at low cost No specifi c cell targeting Specifi c siRNAs reduce infl ammation-associated oligonucleotides (siRNA, Smaller molecules that can lung injury in humans [33] and in animal models shRNA, decoy easily enter cells [28,34] oligonucleotides) Target regulation of specifi c genes shRNA-based approaches have reduced lung injury in animal models [35,36] Cell-delivered gene therapy humans [33] and in animal models [28,34]. shRNA is a single strand of RNA that, when introduced into the cell, is reverse transcribed and integrated into the genome, becoming heritable. During subsequent transcription, the sequence generates an oligonucleotide with a tight hairpin turn that is processed into siRNA. shRNAs have reduced lung injury in animal models [35,36]. Decoy oligonucleotides are double-stranded DNA molecules of 20 to 28 nucleo tides, which bind to specifi c transcription factors to reduce expression of targeted genes, and have been successfully used in animal models [37,38].

Cell-delivered gene therapy
An alternative approach is to use systemically delivered cells to deliver genes to the lung. Th is approach has been used to enhance the therapeutic potential of stem cellssuch as mesenchymal stem/stromal cells, which demon strate promise in preclinical ALI/ARDS models [39]. Fibroblasts have also been used to successfully deliver genes to the lung to attenuate ALI [40]. Preliminary data from a clinical trial in pulmonary hypertension show that endothelial progenitor cells overexpressing endothelial nitric oxide synthase (NOS3) decrease pulmonary vascular resistance [41], highlighting the potential of cell-delivered gene therapy for ALI/ARDS.

Delivery of vector to the lung Airway delivery
Nebulization of genetic material into the lung is eff ective [42], safe and well tolerated [32,43,44]. Th e integrity of AAV vectors [9,43] and adenoviral virus vectors [44] are maintained post nebulization, as are cationic lipid vectors [32] and DNA and RNA oligonucleotides [45]. A number of gene therapy clinical trials have utilized nebulization to deliver the transgene to the lung [23,43], but without clear clinical benefi t to date [43,44].

Intravascular delivery
Intravascular delivery approaches target the lung endothelium. Th ese approaches have been successfully used in preclinical studies of cell-based gene therapies [39,40], and also with vectors that incorporate components such as antibodies to target antigens on the lung endothelium [10].

Barriers to eff ective gene therapy in ARDS
Successful gene-based therapies require the delivery of high quantities of the gene or oligonucleotide to the pulmonary epithelial or endothelial surface, require effi cient entry into the cytoplasm of these large and insoluble nucleic acids, which then have to move from the cytoplasm into the nucleus, and activate transcription of its product. Multiple barriers exist that hinder this process, not least the natural defense mechanisms of the lung, and additional diffi culties that exist in transducing the acutely injured lung (Table 2). Limitations regarding delivery technologies and defi ciencies in our knowledge regarding the optimal molecular targets also reduce the effi cacy of these approaches.

Pulmonary defense mechanisms
Th e lung has evolved eff ective barriers to prevent the uptake of any inhaled foreign particles [46]. While advantageous in minimizing the potential for uptake of external genetic material (for example, viral DNA), these barriers make it more diffi cult to use gene-based therapies in the lung. Barriers to entry of foreign genetic material into the lung include airway mucus and the epithelial lining fl uid, which traps and clears inhaled material. Th e glycocalyceal barrier hinders contact with the cell membrane, while the tight intercellular epithelial junctions and limited luminal endocytosis further restrict entry of foreign material into the epithelial cells.

Diffi culties transducing the injured lung
Transducing the acutely injured lung may be diffi cult, due to the presence of pulmonary edema, consolidated or collapsed alveoli, and additional extracellular barriers such as mucus. Gene-based therapies targeted at the pulmonary epithelium may be less eff ective where there is extensive denudation of the pulmonary epithelium, as may occur in primary ARDS. Encouragingly, there is some evidence to suggest that ALI may not substantially impair viral gene transfer to the alveolar epithelium [47].

Limitations of vector systems
Th e key limitation of nonviral vector approaches has been their lack of effi ciency in mediating gene transfer and transgene expression in the airway epithelium. Viral vectors are immunogenic, due to the protein coat of the viral vector, and the immune response is related to both vector dose and number of administrations. Th e potential to limit administration to a single dose in ALI/ARDS may reduce this risk. However, the development of an infl amma tory response resulting in death following administration of a fi rst-generation adenoviral vector highlights the risks involved [48]. Additional limitations of viral vectors include transgene size, which is limited by the size of the capsid that encloses the viral genes.

Insights from preclinical ALI/ARDS studies
Th e therapeutic potential of gene therapy for ALI/ARDS is underlined by a growing body of literature demon strating effi cacy in relevant preclinical models. In considering the clinical implications of these studies, it is important to acknowledge that animal models of ARDS do not fully replicate the complex pathophysiological changes seen in the clinical setting. Th is is highlighted by the fact that many pharmacologic strategies demonstrating considerable promise in preclinical studies were later proven ineff ective in clinical trials. Nevertheless, these studies provide insights into the clinical potential of these strategies.
In contrast, overexpression of IL-1β can directly cause ALI [56], while overexpression of suppressor of cytokine signal ing-3 worsens immune-complex-induced ALI [57]. Intriguingly, intra tracheal administration of adenoviral vector incor porating IL-10, prior to zymosan-induced lung injury, improved survival at a lower dose but was ineff ective and even harmful at higher doses [58].
More recently, aerosolization of siRNA that targets respiratory syncytial virus viral replication was safe and potentially eff ective in patients post lung transplant with respiratory syncytial virus infection [33], clearly illustrating the therapeutic potential of these approaches for ALI/ARDS.

Studies using cell-delivered gene therapy
Mei and colleagues enhanced the effi cacy of mesen chymal stem/stromal cells in endotoxin-induced ALI by transducing them to overexpress angiopoeitin-1 (Figure 4) [39]. Mesenchymal stem/stromal cells overexpressing IL-10 decreased alveolar infi ltration of CD4 and CD8 T cells following lung ischemia-reperfusion injury [65]. Bone marrow stem cells expressing keratinocyte growth factor attenuate bleomycin-induced lung injury [66]. Non stem cells can also be used to deliver genes to the injured lung [67]. Fibroblasts overexpressing angiopoeitin-1 attenuate endotoxin-induced lung injury [40], while fi broblasts overexpressing vascular endothelial growth factor and endothelial nitric oxide synthase can attenuate or even reverse endotoxin-induced ALI [68].

Gene-based therapies for ALI/ARDS: future directions
Advances in the identifi cation of therapeutic targets, improvements in viral and nonviral vector technologies, and regulation of gene-based therapies by temporal and spatial targeting off er the potential to translate the therapeutic promise of gene-based therapies for ALI/ ARDS to the clinical setting (Table 3).

Better viral vectors
Viral vectors remain the focus of intensive research to optimize their effi ciency, to minimize their immuno genicity and to enhance their tissue specifi city [19,31,69,70]. Strategies to develop less immunogenic vectors have focused on modifying the naturally occurring proteins in the viral coat [71]. Much research has been devoted to searching and characterizing both naturally occurring [71] and engineered capsid variants from mammalian species [72]. Capsid protein modification has also been used to enhance tissue specifi city [70]. Envelope protein pseudotyping involves encapsulating the modifi ed genome from one virus, such as simian immuno defi ci ency virus, with envelope proteins from another virus, such as vesicular stomatitic virus. Th is encapsu lation can enhance the therapeutic potential of viral vectors, by combining the advantages of one viral genome (for example, bigger payload or site-specifi c integration) with the tissue tropism of another virus.

Better nonviral vectors
Strategies to enhance the eff ectiveness of the lipoplexes used to deliver plasmids and other DNA/RNA oligonucleotides involve manipulation of the lipoplex lipid content and the use of targeting peptides. Th e choice of lipid infl uences expression effi ciency by enhancing release of the genetic material within the target cell [73,74]. Targeting peptides increases transfection effi ciency by directing the lipid to a particular cell membrane or cell type [31]. Physical methods of plasmid delivery such as electroporation [60] and ultrasound can enhance gene transfer by bringing the plasmid DNA into closer proximity with the cell membrane and/or causing temporary disruption of the cell membrane. Other physical methods can also be used to increase in vivo gene transfer, including pressurized vascular delivery, laser, magnetic fi elds and gene gun delivery. Th ese systems enable plasmid-based gene delivery to reach effi ciencies close to that achieved with viral vectors.

Enhanced gene expression strategies
Successful gene therapy relies upon being able to target the injury site, and to control the duration and levels of gene expression. Modifying the transgene DNA to exclude nonmethylated CpG motifs, typical of bacterial DNA, decreases the immune response and may increase transgene expression [75,76]. High-effi ciency tissue-specifi c promoters may improve the effi ciency and specifi city of transgene expression. Lung-specifi c promoters include surfactant promoters [77] such as the surfactant protein C promoter [78], a ciliated cell-specifi c promoter FOXJ1 [79], the cytokeratin 18 promoter [80], and the Clara cell 10-kDa protein [78]. Promoters can also be used to target a specifi c phase of illness, switching on when required to produce an eff ect at the optimal time point. A related approach is the development of promoters that allow for transfected genes to be turned on and off . Currently, the tetracycline-dependent gene expression vector [81] is the most widely used regulated system as it has a good safety profi le. Tetracycline is rapidly metabolized and cleared from the body, making it an ideal drug to control gene expression. However, the potential for an activator such as tetracycline to modulate the lung injury should be borne in mind. New-generation transactivators, with no basal activity and increased sensitivity, have now been developed [82]. In an ARDS context, conditional regulation of gene expression by the combined use of a lung-specifi c promoter and the tetracycline-dependent gene expression system may be a useful approach [83].  Capsid protein modifi cation to reduce immunogenicity [71] Capsid protein modifi cation to enhance tissue specifi city [70] Envelope protein pseudotyping

Nonviral vectors
Manipulation of lipoplex lipid content to enhance cellular uptake [73,74] Use of targeting peptides on lipoplexes and polyplexes [31] Strategies to enhance gene transfer; for example, electroporation, ultrasound, gene gun delivery

Gene expression strategies
Modifying transgene DNA to eliminate bacterial motifs [75,76] Development of high-effi ciency tissue-specifi c promoters [77][78][79][80] Development of promoters that regulate gene expression [83] Enhanced therapeutic targeting Nebulization technologies [9] Strategies to target the pulmonary endothelium [10] Improved cellular uptake of vector Surface active agents to enhance vector spread [84] Reduce ubiquitination of viral capsid proteins [85] Better therapeutic targets Enhancement or restoration of lung epithelial and/or endothelial cell function [86] Strengthening lung defense mechanisms against injury [87] Speeding clearance of infl ammation and infection Enhancement of the repair process following ALI/ARDS [88].

Enhanced therapeutic targeting
An advantage of gene-based strategies is the ability to target specifi c cells within an organ; for example, the epithelial cells of the lung. Novel nebulization technologies, which facilitate the delivery of large quantities of undamaged vector to the distal lung, demonstrate considerable promise in this regard [9]. Alternative approaches to spatial targeting include targeting specifi c receptors that are plentiful on the target cell to increase transfection effi ciency. An interesting development in this regard is the targeting of systemically administered therapies to the pulmonary endothelium using antibodies to proteins expressed preferentially on these cells ( Figure 5) [10]. In these studies, the antioxidant enzyme catalase was conjugated with antibodies to the adhesion molecule PECAM, which is widely expressed on pulmonary endothelial cells, and to a nonspecifi c IgG antibody. Th e anti-PECAM/catalase conjugate, but not the IgG/catalase conjugate, bound specifi cally to the pulmonary endothelium and attenuated hydrogen peroxide injury.

Improved cellular uptake of vector
Specifi c strategies have been developed to maximize uptake of vector into alveolar epithelial cells. It is possible to enhance lung transgene expression with the use of surface-active agents such as perfl urocarbon, which enhances the spread of vector and mixing within the epithelial lining fl uid [84]. Agents that reduce ubiquitination of AAV capsid proteins following endocytosis, such as tripeptide proteasome inhibitors, dramatically augment (>2,000-fold) AAV vector transduction in airway epithelia [85].

Better therapeutic targets
Ultimately, the success or failure of gene-based therapies for ALI/ARDS is likely to rest on the identifi cation of better gene targets. Ongoing advances in our understanding of the pathophysiology of ALI/ARDS continue to reveal novel therapeutic targets for gene-based approaches. Promising potential approaches include strate gies to enhance or restore lung epithelial and/or endothelial cell function [86], to strengthen lung defense mechanisms against injury [87], to speed clear ance of infl ammation and infection, and to enhance the repair process following ALI/ARDS [88].

Summary and conclusions
ALI/ARDS may be a particularly suitable disease process for gene-based therapies (Table 4). Th is is supported by increasing evidence from relevant preclinical ARDS models for the effi cacy of gene-based therapies that enhance or restore lung epithelial and/or endothelial cell function, strengthen lung defense mecha nisms against injury, speed resolution of infl ammation and infection, and enhance the repair process following ALI/ARDS. Despite this promising preclinical evidence, the potential for gene based approaches to ALI/ARDS in the clinical setting remains to be realized. Multiple barriers exist to the successful use of gene-based therapies in the lung, which limit the effi cacy of these approaches. Future research approaches should focus on overcoming these barriers, by developing more eff ective and less immunogenic vector delivery systems, developing strategies to focus gene expression on specifi c injury zones of the lung for defi ned time periods, and identifying better molecular targets that can take advantage of these potentially very powerful therapeutic approaches.

Competing interests
The authors declare that they have no competing interests.