Acute Lung Injury/acute Respiratory Distress Syndrome -a Therapeutic Challenge Clinical Review: Stem Cell Therapies for Acute Lung Injury/acute Respiratory Distress Syndrome -hope or Hype?

Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are the leading cause of death in critical care, with mortality rates of 40 to 60%. In the US alone, there are 200,000 new cases annually [1]. ALI/ ARDS also constitute a considerable long-term illness and disability burden, with signifi cant neuromuscular, pulmonary and psychological morbidity seen in 50 to 70% of survivors, and just 49% returning to employment one year post-discharge [2]. Despite being a focus of ongoing intensive research eff orts over four decades, there are no pharmacologic therapies for ALI/ARDS [3]. Large-scale clinical trials of multiple therapeutic strategies, including nitric oxide [4,5], anti-oxidants [6-9], surfact-ants [10], corticosteroids [11] and immunomodulating agents such as IL-10 [12] and granulocyte-macrophage colony-stimulating factor [13] have all failed. Consequently , advances in the management of ALI/ARDS have relied on improvements in supportive measures, such as 'protective' mechanical ventilation strategies [3], restrictive intravenous fl uid management approaches [14], and prone positioning of severely hypoxaemic patients [15,16]. While these and other improvements in supportive care have decreased mortality [17], the failure of pharmacologic therapies suggests the need to consider novel approaches for ALI/ARDS. ALI/ARDS is a highly complex disease process. Earlier concepts of distinct disease phases, from an early 'pro-infl ammatory' to a later 'fi brotic' phase, now appear to be an over-simplifi cation. Th ese 'phases' largely co-exist, with evidence of pro-infl ammatory responses leading to host damage, an impaired immune response to pathogens , and repair and fi brosis all present in the complex milieu that is clinical ALI/ARDS. Given this, it is perhaps not surprising that strategies targeted at one aspect of the disease process have been unsuccessful. Th is suggests the need to consider more complex therapeutic approaches aimed at reducing early injury while maintaining host immune competence, and facilitating (or at least not inhibiting) lung regeneration Abstract A growing understanding of the complexity of the pathophysiology of acute lung injury (ALI)/acute respiratory distress syndrome (ARDS), coupled with advances in stem cell biology, has led to a renewed interest in the therapeutic potential of stem cells for this devastating disease. Mesenchymal stem cells appear closest to clinical translation, given the evidence that they may favourably modulate the immune response to reduce lung injury, while maintaining host immune-competence and also facilitating lung regeneration and repair. The demonstration that human mesenchymal stem cells exert benefi t in the endotoxin-injured human lung is …

Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are the leading cause of death in critical care, with mortality rates of 40 to 60%. In the US alone, there are 200,000 new cases annually [1]. ALI/ ARDS also constitute a considerable long-term illness and disability burden, with signifi cant neuromuscular, pulmonary and psychological morbidity seen in 50 to 70% of survivors, and just 49% returning to employment one year post-discharge [2]. Despite being a focus of ongoing intensive research eff orts over four decades, there are no pharmacologic therapies for ALI/ARDS [3]. Largescale clinical trials of multiple therapeutic strategies, including nitric oxide [4,5], anti-oxidants [6][7][8][9], surfactants [10], corticosteroids [11] and immunomodulating agents such as IL-10 [12] and granulocyte-macrophage colony-stimulating factor [13] have all failed. Consequently, advances in the management of ALI/ARDS have relied on improvements in supportive measures, such as 'protective' mechanical ventilation strategies [3], restrictive intravenous fl uid management approaches [14], and prone positioning of severely hypoxaemic patients [15,16]. While these and other improvements in supportive care have decreased mortality [17], the failure of pharmacologic therapies suggests the need to consider novel approaches for ALI/ARDS. ALI/ARDS is a highly complex disease process. Earlier concepts of distinct disease phases, from an early 'proinfl ammatory' to a later 'fi brotic' phase, now appear to be an over-simplifi cation. Th ese 'phases' largely co-exist, with evidence of pro-infl ammatory responses leading to host damage, an impaired immune response to pathogens, and repair and fi brosis all present in the complex milieu that is clinical ALI/ARDS. Given this, it is perhaps not surprising that strategies targeted at one aspect of the disease process have been unsuccessful. Th is suggests the need to consider more complex therapeutic approaches aimed at reducing early injury while maintaining host immune competence, and facilitating (or at least not inhibiting) lung regeneration Abstract A growing understanding of the complexity of the pathophysiology of acute lung injury (ALI)/acute respiratory distress syndrome (ARDS), coupled with advances in stem cell biology, has led to a renewed interest in the therapeutic potential of stem cells for this devastating disease. Mesenchymal stem cells appear closest to clinical translation, given the evidence that they may favourably modulate the immune response to reduce lung injury, while maintaining host immune-competence and also facilitating lung regeneration and repair. The demonstration that human mesenchymal stem cells exert benefi t in the endotoxin-injured human lung is particularly persuasive. Endothelial progenitor cells also demonstrate promise in reducing endothelial damage, which is a key pathophysiological feature of ALI. Embryonic and induced pluripotent stem cells are at an earlier stage in the translational process, but off er the hope of directly replacing injured lung tissue. The lung itself also contains endogenous stem cells, which may ultimately off er the greatest hope for lung diseases, given their physiologic role in replacing and regenerating native lung tissues. However, signifi cant defi cits remain in our knowledge regarding the mechanisms of action of stem cells, their effi cacy in relevant pre-clinical models, and their safety, particularly in critically ill patients. These gaps need to be addressed before the enormous therapeutic potential of stem cells for ALI/ARDS can be realised. and repair. Could stem cells fi t this new therapeutic paradigm?

What are stem cells?
A stem cell is a cell that has the ability to divide asymmetrically to produce another cell like itself, or a more diff erentiated cell ( Figure 1). Stem cells are classifi ed based on their tissue of origin. Embryonic stem cells (ESCs) are derived from the inner blastocyst cell mass ( Figure 2) and are pluripotent, that is, they are capable of diff erentiating into cells of all embryological lineages. Fetal stem cells are derived from extra-embryonic tissues, including the amniotic fl uid, the placenta, umbilical cord blood and Wharton's jelly. Th ey appear to represent a new stem cell population with growth kinetics and plasticity intermediate between adult and embryonic stem cells.
Adult tissue-derived stem cells include mesenchymal stem cells (MSCs), endothelial progenitor cells (EPCs), and endogenous lung stem cells ( Figure 2). MSCs are cells of stromal origin that are capable of self-renewal and diff erentiation into cells of mesodermal origin, including chondrocytes, osteocytes and adipocytes [18]. Endothelial progenitor cells are circulating cells with the ability to proliferate and diff erentiate into mature endothelial cells. Th e adult lung itself also contains endogenous stem cells ( Figure 3). Th ese lung stem cells may ultimately be the ideal cell to regenerate injured lung. In general, the potential of stem cells to diff erentiate decreases as one moves from embryonic to adult stem cells. Adult stem cells are multi-potent, having the potential to diff erentiate into a more limited range of mature cell types. An exception is induced pluripotent stem cells (iPSCs), derived from adult cells that have been reprogrammed to de-diff erentiate following transduction with transcription factors [19].

Why do stem cells off er promise for ALI/ARDS?
Stem cells off er considerable promise as a novel therapeutic strategy for ALI/ARDS for a number of reasons. First, stem cells, particularly pluripotent cells such as ESCs or iPSCs, off er the potential to diff erentiate into lung cells and directly replace damaged cells and tissues. Second, ALI is characterized by an intense but transient infl ammatory response. While previous strategies to simply inhibit this response have met with failure, the more complex, 'immunomodulatory' properties of adult stem cells such as MSCs may be more eff ective. Stem cells may be able to 'reprogramme' the immune response to reduce the destructive infl ammatory elements while preserving the host response to pathogens. Th ird, stem cells may be able to enhance the repair and resolution of lung injury. Resolution of ALI/ARDS is impeded by destruction of the integrity of the epithelial barrier, which inhibits alveolar fl uid clearance and depletes surfactant [20]. Stem cells may restore epithelial and endothelial function, whether by diff erentiating into these cell types or via secretion of paracrine factors to enhance restoration of these tissues. Fourth, ALI/ARDS is frequently a component of a generalized process resulting in dysfunction and failure of multiple organs. MSCs have been demonstrated to decrease injury and/or restore function in the kidney [21,22], liver [23,24] and heart [25]. Fifth, stem cells may directly attenuate bacterial sepsis, the commonest [26] and most severe [27] cause of ALI/ARDS, via a number of mechanisms, including enhancement of phagocytosis, increased bacterial clearance [28], and anti-microbial peptide secretion [29]. Sixth, there is the potential to further enhance the therapeutic eff ect of MSCs by tranducing them to secrete disease modifying molecules. As stem cells home to sites of infl ammation when administered intravenously follow ing tissue injury [30], they may therefore provide an attractive vector for gene-based therapies [31]. Seventh, both the distal lung epithelium and the pulmonary endothelium are selectively accessible to stem cell therapies, via the intratracheal route [32] and because the entire cardiac output transits the pulmonary vasculature. Finally, stem cells are in clinical studies for a wide range of disease processes. Th e clinical potential of MSCs for ALI/ARDS has been considerably enhanced by a recent study demonstrating that human MSCs can reduce endotoxin-induced injury to explanted human lungs [33].
Taken together, these fi ndings off er considerable hope for stem cells as a therapy for ALI/ARDS. However, the potential mechanisms of action, therapeutic eff ects and translational potential diff er signifi cantly based on the specifi c stem cell population. While MSCs currently may off er the best hope of a potential therapy for ALI/ARDS [33], each stem cell population is a potential therapeutic candidate (Table 1). In the following paragraphs we consider each population separately.

How do stem cells work?
Stem cells may act directly via diff erentiation into one or more of the injured cell types and engraftment into the organ. Indirect mechanisms of action include modulation of the injury and repair process via paracrine and cell-cell contact-dependent mechanisms. Th e eff ects of specifi c stem cells appear to diff er depending on the stem cell type.

Diff erentiation and engraftment
Cellular diff erentiation is the process by which a less specialized cell becomes a more specialized cell type. Pluripotent ESCs or iPSCs have the capacity to diff erentiate into cells of all three dermal layers, that is, endoderm, mesoderm and ectoderm ( Figure 2). Diff erentia tion into the injured cell types and engraftment appears to be a key mechanism of action for these stem cell types. Th e derivation of alveolar epithelial cells from ESCs has been demonstrated [34].
Th e role of diff erentiation and engraftment in the mechanisms of action of adult stem cells is more controversial. Adult stem cells have a more restricted ability to diff erentiate, being committed to a specifi c lineage ( Figure 2). Multipotent adult stem cells such as MSCs can diff erentiate into a variety of 'lineage-specifi c' (that is, mesenchymal) tissues, including bone, cartilage, tendon, fat, bone marrow stroma and muscle [35]. Progenitor cells, such as EPCs, have even more limited diff erentiation potential, being committed to a single lineage. Trans-diff erentiation refers to the potential for adult stem cells to diff erentiate into cells of other lineages [36]. Early studies suggested trans-diff erentiation might constitute an important mechanism of action of adult stem and progenitor cells. Krause and colleagues [37] found that a single bone marrow-derived haemato poietic stem cell could give rise to cells of multiple diff er ent organs, including the lung. Kotton and colleagues [38] demonstrated that bone marrow-derived cells could engraft into pulmonary epithelium and exhibit character istics specifi c to lung epithelial cells ( Figure 4). Suratt and colleagues [39] found signifi cant rates of engraftment of transplanted haematopoietic stem cells in lung specimens from female allogeneic haematopoietic stem cell transplant recipients that received stem cells from male donors. However, more recent pre-clinical studies demonstrate that while adult stem cells can reduce ALI/ARDS, engraftment rates are low [31,40,41], casting doubt on the therapeutic signifi cance of this mechanism of action.

Paracrine and cell-cell eff ects
Th e potential for stem cells to act indirectly, by modulating key aspects of the host immune and repair responses, while potentially applicable to all stem cells types, is best understood for MSCs. MSCs secrete multiple paracrine factors, such as growth factors [42], factors regulating endothelial and epithelial permeability [43], anti-infl ammatory cytokines [44], and antimicrobial peptides [29], that can modulate the immune response and facilitate repair and regeneration ( Figure 5). Cell-cell contact, mediated by adhesion molecules, appears to be a key mechanism by which MSCs modulate immune eff ector cells such as macrophages [44] and T cells [45]. MSCs also appear to communicate with host cells by releasing membrane-derived microvesicles containing RNA [45], and mitochondria [46].

Therapeutic potential of stem cells for ALI/ARDS
A growing body of evidence highlights the potential benefi ts of cell-based therapy for ALI/ARDS. Most pre-clinical studies to date have focused on MSCs, and the therapeutic potential of these cells is examined in detail elsewhere [47]. However, emerging studies using ESCs, EPCs and foetal stem cells attest to the growing interest in a variety of cell-based approaches for ALI/ ARDS.

Mesenchymal stem cells
MSCs were fi rst identifi ed in the bone marrow in 1976 by Friendenstein and colleagues [48], but have now been identifi ed in numerous tissues, including lung, umbilical cord, cord blood, adipose tissue and gastrointestinal tract. MSCs exhibit several advantages of relevance to ALI/ARDS, including their derivation from (multiple) adult tissues, their low immunogenicity, which means that they may be given allogeneically [49], and their relative ease of isolation and enormous expansion potential in culture.

Effi cacy in pre-clinical models
MSCs have been demonstrated to reduce the severity of lung injury caused by septic and non-septic insults. Ortiz and colleagues [50] reported that MSC therapy decreased bleomycin-induced lung injury in mice despite an engraftment rate of less than 5%. MSCs and/or their secreted factors reduced hyperoxia-induced lung infl ammation, decreased histologic injury, and attenuated longterm remodelling [51,52]. Most recently, they have been demonstrated to enhance recovery and repair [53] in a model of repair following ventilation-induced lung injury [54]. Importantly, MSCs appear to enhance the host response to bacterial sepsis; they decrease septic lung injury induced by endotoxin [40,46], Escherichia coli pneu monia [28] and systemic sepsis [46]. Intravenous MSC administration decreased intraperitoneal endotoxininduced alveolar neutrophil infi ltration and decreased pulmonary edema [55]. MSC therapy decreased the number of bacteria recovered from the lung following E. coli instillation [56] and following cecal ligation and puncture [56]. MSCs improved survival, reduced organ dysfunction, including indices of ALI, reduced neutrophil oxidative injury and increased circulating neutrophils, while lowering bacterial counts in blood, following cecal ligation and puncture-induced systemic sepsis [44]. Th e anti-bacterial eff ects of MSCs appear to be due, in part, to the secretion of anti-microbial peptides [29].
Critical illness is characterised by dysfunction and failure of multiple organs. Reassuringly, MSCs have demon strated effi cacy in pre-clinical studies for a variety of clinical disorders, including myocardial infarction [33,57], diabetes [58], hepatic failure [59], acute renal failure [60], and sepsis [28,44]. Th e clinical potential of MSCs for ALI/ARDS has been considerably enhanced by a recent study demonstrating that human MSCs can In the trachea and main bronchi, undiff erentiated basal cells (stained for transcription factor p63) can function as classical stem cells. Duct cells from submucosal glands located in the cartilaginous airways are also a potential niche. In the more distal lung, Clara cells and 'variant' Clara cells (stained red) are found in the bronchioles and bronchiolar-alveolar junctions, respectively, while the alveoli contain type II cells (stained green) that can regenerate type 1 pneumocytes. Reproduced with permission from [105].
reduce endotoxin-induced injury to explanted human lungs [61] (Figure 6). Taken together, these fi ndings off er considerable hope for MSCs as a therapy for ALI/ARDS

Insights from clinical studies
MSCs are in clinical trials for multiple other diseases. In fact, the Clinicaltrials.gov database has over 100 clinical trials on its registry investigating MSCs as a therapy for diverse diseases, including diabetes, myocardial infarction, Crohn's disease, graft versus host disease, osteogenesis imperfecta and multiple sclerosis. A pilot study with 11 patients showed that MSCs were eff ective for radiation-induced lung injuries that developed after combined chemotherapy and radiation therapy for lympho granulomatosis or breast cancer [62]. In a phase I study of MSCs (Prochymal®, Osiris Th erapeutics Inc., Columbia, MD, US) in patients with acute myocardial infarction, MSCs improved forced expiratory volume in 1 s (FEV 1 ) and forced vital capacity (FVC) [63]. Th ese observations stimulated a multicenter phase II trial of MSCs for patients with moderate to severe chronic obstructive pulmonary disease (NCT00683722). Recent clinical studies highlight the potential for MSCs to enhance wound repair. Treatment of cutaneous wounds with MSCs accelerates wound healing kinetics and increases epithelialization and angiogenesis [64][65][66]. Th ese clinical trials attest to the safety of MSCs and their ability to attenuate injury severity whilst enhancing repair.

Mechanisms of action
Th e mechanism of action of MSCs appears to be predominantly paracrine, and to involve the release of factors that have immunomodulatory, reparative and anti-bacterial eff ects. In contrast to classic 'antiinfl ammatory' strategies, MSCs decrease host damage arising from the infl ammatory response while enhancing host resistance to sepsis. MSCs interact with a wide range of immune cells and exert diverse eff ects on the innate and adaptive immune responses. Th ese include suppression of T-cell proliferation, natural killer cell function and inhibition of dendritic cell diff erentiation [67]. Soluble factors that are candidate mediators for MSC immune modulation include transforming growth factor-β [68], indoleamine 2,3-dioxygenase [69], IL-1-receptor antagonist [41], tumour necrosis factor-α-induced protein (TSG)-6 [46], and IL-10 and prostaglandin E2 [44] ( Figure 5). MSCs appear to augment the host immune response to sepsis via secretion of anti-microbial peptides and TSG-6, and they increased bacterial clearance and enhanced host cell phagocytosis in septic mice [28].
Recently, Krasnodembskaya and colleagues [29] reported that MSCs improved bacterial clearance of E. coli pneumonia via the secretion of an antimicrobial peptide, LL-37.
MSCs appear to aid lung repair and regeneration following injury, in part via the secretion of cyto protective agents [31,33,43,46]. MSC secretion of angiopoietin and keratinocyte growth factor restores alveolar epithelial and endothelial permeability and enhances resolution of ALI/ARDS in pre-clinical models [31,33,43,46]. MSCs decreased bleomycin-induced lung injury and fi brosis, and decreased lung collagen accumulation, fi brosis and levels of matrix metalloproteinases in part by IL-1receptor antagonist secretion [41].

Barriers to clinical translation
Th e optimal route of administration of MSCs is not known, with evidence supporting the intravenous [44], intratracheal [33,40] and intraperitoneal administration routes [46]. Th e optimal dosage regimen for MSCs, including the lower eff ective MSC doses, is also unclear. Th e true 'therapeutic' potential of MSCs -that is, their eff ectiveness when administered after the lung injury is established -is also unclear. Pre-clinical studies to date have used relatively poorly defi ned, heterogeneous MSCs. Th e potential for recently identifi ed more specifi c MSC subpopulations [58,70] to be more eff ective in attenuating ALI/ARDS remains to be determined. Unlike haematopoietic stem cells, MSCs are not defi ned by a single marker and stem cell markers are not uniquely expressed by stem cells [18,70]. While a set of minimal criteria for defi ning MSCs has been developed [71], there remains a lack of standardised protocols for isolation and characterisation of them. Furthermore, there is no validated method of measuring MSC bioactivity in vivo [72]. Despite recent advances, our understanding of the mechanisms of action of MSCs remains incomplete. Of concern in this regard, recent studies suggest that MSCs might elicit a memory T-cell response in mice, suggesting that they are not as 'immunoprivileged' as previously thought [73,74]. Th e need to address these barriers to clinical translation is underlined by the limited clinical experience with MSCs in critically ill patients to date [63].

Endothelial progenitor cells
Interest in EPCs stems from the fact that endothelial damage is a key pathophysiological feature of ALI, contributing to hyaline membrane formation and development of protein-rich alveolar edema [20]. While there is also no defi ned set of cell markers to identify an EPC, two diff erent subpopulations of EPCs, with diff erent diff erentiation abilities, cell markers and roles in endothelial repair, have been identifi ed. 'Early' EPCs have haematopoietic surface markers (for example, CD-34) and secrete pro-angiogenic factors but display limited diff erentiation ability. 'Late' EPCs, so called because they appear after more than two weeks in culture, lack Photomicrographs of murine lung demonstrating engraftment of mesenchymal stem cells, which are stained blue due to their expression of the Lac-Z transgene, into the mouse lung following bleomycin-induced lung injury. Reproduced with permission from [38].
haematopoietic surface markers and do not secrete proangiogenic factors. However, they make endothelioid tubes in vitro and have a greater role in replacement of damaged endothelium [75].

Effi cacy in pre-clinical models
EPCs appear to regenerate the alveolar endothelium in ALI. Lam and colleagues [76] demonstrated that autologous infusion of 'early' EPCs improved alveolar-capillary integrity as evidenced by decreased lung haemorrhage, water content and hyaline membrane formation in a rabbit oleic acid ARDS/ALI model. EPCs may also exert anti-infl ammatory as well as regenerative eff ects [56]. Animals treated with allogeneic EPCs showed increased levels of the anti-infl ammatory cytokine IL-10 as well as reduced inducible nitric oxide synthase and endothelin-1 [56]. Th e EPC group also showed evidence of neovasculari sation, re-endothelialization and improved survival. Importantly, this study was carried out using allogenic EPCs and attests to their low immunogenicity [56].

Insights from clinical studies
Human studies showing an association between increased circulating levels of EPCs and improved outcome from ALI [77] and bacterial pneumonia [78] suggest a role for EPCs in lung repair. Two clinical trials investigating the use of autologous EPCs in pulmonary hypertension have been conducted. A pilot study in children [79] and a small randomised control trial in adults [80] showed promising results with improvements in exercise capacity and pulmonary haemodynamics in the treatment group. No immunologic reactions or other adverse eff ects were noted from EPC infusion. A comparable trial of administration of autologous EPCs transduced to express endothelial nitric oxide synthase for adult patients with pulmonary hypertension is currently underway in Canada.

Mechanisms of action
EPCs appear to exert therapeutic eff ects via diverse direct diff erentiation and engraftment into the vasculature and via secretion of factors that mobilise adjacent endothelial cells and tissue-resident progenitor cells to partake in angiogenesis and reconstruction. Th e EPC secretome has recently been analyzed and found to contain at least 35 proteins known to relate to endothelial cell biology and angiogenesis [81]. As discussed above, EPCs also appear to exert immunomodulatory eff ects.

Barriers to clinical translation
Low circulating levels of EPCs, even in illness, mean that these cells are diffi cult to isolate [82]. Most studies use autologously harvested EPCs. Experience with allogeneic transplantation of EPCs is limited [56] and the safety of this approach is unconfi rmed. Th is is a signifi cant practical limitation that reduces their therapeutic potential for acute illnesses such as ALI/ARDS.

Embryonic stem cells
ESCs are pluripotent cells derived from the inner blastocyst cell mass (Figure 2) and constitute a potentially unlimited source of cells that could be diff erentiated into lung progenitor cells for possible clinical use. Th e generation and implantation of a committed lung progeni tor cell would have a clear therapeutic advantage, as it could give rise to all cell types within the lung following therapeutic transplantation. However, the embryologic origin of lung cells makes the generation of lung cells from ESCs diffi cult, and there are signifi cant ethical issues regarding the use of these cells.

Insights from laboratory studies
Advances in ESC culture and diff erentiation methods in recent years have led to renewed interest in ESCs as a potential therapeutic agent in ALI [83,84]. Eff orts have largely concentrated on derivation of type II alveolar epithelial cells from ESCs [34,85]. A diffi culty is the origin of the lung in the endoderm, the third germ layer to form and one with a complex cellular specifi cation process. Consequently, alveolar epithelial cells have proven diffi cult to derive from both murine and human ESCs. Wang and colleagues [34] recently derived a transfection and culture procedure utilising surfactant protein C promoter-driven neomycin expression to facilitate the diff eren tiation of human ESCs into a more than 99% pure population of type II alveolar epithelial cells. Previous techniques, including co-culture [86] with pulmonary mesen chyme, yield mixed populations of cell derivatives and were unsuitable for clinical use. Purity of cell popu lations for transplantation is essential as the presence of pluripotent cells in the diff erentiating culture engenders a signifi cant risk of teratoma formation post-transplantation.

Effi cacy in pre-clinical models
In vivo studies of ESCs as a potential replacement therapy for damaged alveolar epithelial cells in ALI remain limited [87,88]. Th e functional signifi cance of ESC diff eren tiation and engraftment into the injured lung remains to be clearly demonstrated [88]. Roszell and colleagues [85] developed alveolar type II (ATII) cells from mouse ESCs and administered them intratracheally into preterm mice. Over the 24-hour study period the cells maintained diff erentiation and surfactant protein C expression. However, it was not determined if the cells had engrafted into the lung parenchyma or resulted in any functional benefi t. Wang and colleagues investigated the eff ect of ESC-derived ATII cells on ALI in mice injured by intratracheal bleomycin [88]. At day 9 post-intratracheal administration, these human ESC-ATII cells had engrafted in the lung, and a small number had undergone diff erentiation to type I alveolar epithelial cells. Furthermore, the severity of ALI was reduced and animal survival increased. Most recently, Toya and colleagues [89] studied the eff ects of blast progenitor cells derived from human ESCs cultured in conditions favouring development of mesoderm in a mouse caecal ligation and puncture model. Th ese cells ameliorated sepsis-induced lung injury when administered intravenously one hour post-injury by modulation of the immune response. Th e protective eff ects were mediated by a subpopulation of progenitor cells positive for the endothelial and haematopoietic lineage marker angiotensin converting enzyme. Th ese cells interacted with CD11b+ host immune cells in the lungs, reducing their production of pro-infl ammatory cytokines and nitric oxide.

Barriers to clinical translation
Research in this fi eld has been impeded by ethical, fi nancial and technical concerns. Additional concerns include the potential for malignant transformation (that is, teratoma formation) [90] and immune rejection [91].

Foetal stem cells
Foetal stem cells appear to be a novel stem cell population and are derived from extra-embryonic tissues (amniotic fl uid, the placenta, umbilical cord blood and Wharton's jelly). Th ey exhibit growth kinetics and plasticity inter mediate between adult MSCs and ESCs. Unlike ESCs, they demonstrate low immunogenicity in vivo [92] and do not give rise to tumours. Furthermore, given their role in foetal tolerance, these cells have immunoregulatory properties [92] that are already used clinically in ophthal mology [93] and as a therapy for burns [92]. Th eir derivation from extraembryonic tissues, which are normally discarded at birth, obviates the necessity for invasive biopsy or destruction of the blastocyst, and renders them an ethically neutral source of stem cells. Stem cells with a phenotype consistent with MSCs derived from Wharton's jelly of the human umbilical cord reduce bleomycin-induced lung injury and fi brosis [94].

Barriers to clinical translation
Th ese cells are relatively newly defi ned and poorly characterized. Signifi cant additional pre-clinical study is required prior to consideration of their potential in ALI/ ARDS.

Induced pluripotent stem cells
iPSCs are adult somatic cells (for example, dermal fi broblasts) that have undergone dediff erentiation follow ing re programming following transduction (using retroviruses or non-viral techniques) to express four transcription factors, Ocet3/4, Sox2, Klf-4 and c-Myc [19]. Th e autologous nature of the cells eliminates the problem of immune rejection associated with ESCs. iPSCs are comparable to ESCs in terms of morphology, gene expression and teratoma formation. Th e potential applications of iPSCs in ALI/ARDS are numerous (Figure 7).

Barriers to clinical translation
Currently there are no clinical or pre-clinical studies of the use of iPSCs for any respiratory pathology. Th e developmental potential of human iPSC lines has yet to be characterised. Th e recent demonstration of in vitro diff erentiation of iPSCs into ATII-like epithelial cells is a welcome development [95]. However, there is evidence to suggest that iPSCs are neither as robust nor as pleuripotent as ESCs [96]. Moreover, the current gene transduction techniques for generating pluripotent cells are impractical for producing a clinically reliable stem cell source, especially for an acute illness such as ALI [97].

Endogenous lung stem cells
Th e ideal cell type to use to regenerate the injured lung would be the lung's native stem cell population. Th e ability of the lung to regenerate following injury provides clear evidence for the existence of one or more native lung stem cell populations [98]. Until recently, however, this putative adult lung-derived stem cell population remained poorly characterised. While several diff erent possible populations of anatomically and functionally dis tinct endogenous stem cell candidates have been identifi ed, none of these cells meet the full criteria of what constitutes a stem cell. However, in a landmark study Kajstura and colleagues [99] have recently identifi ed a population of c-kit-positive stem cells in the human lung located in the distal lung, the bronchioles and to a lesser extent the alveoli, demonstrating the ability to regenerate all components of an injured mouse lung ( Figure 8). Th is study has generated controversy, Figure 8. Human lung stem cells regenerate the injured mouse lung. Photomicrographs demonstrating engraftment of human lung c-kitpositive cells into the cryo-injured mouse lung. These cells, which express green fl uorescent protein (EGFP), regenerated the injured area of lung tissue. The c-kit-positive cells diff erentiated into lung epithelial cells (area 1), and formed lung bronchioles as evidenced by staining for Clara cell 10-kDa secretory protein (CC-10), alveolar lung tissue (area 2) as evidenced by staining for aquaporin-5 (AQP5), and pulmonary arterioles (area 3) as evidenced by staining for von-Willebrand factor (VWF). Reproduced with permission from [99]. particu larly in regard to the fact that key experiments lacked appropriate controls, and the fact that, by focusing on human c-kit-positive lung cells, these studies provide only partial insights into the fates of these cells [100,101]. Th e novelty of these fi ndings is underlined by the fact that multipotent stem cells that can give rise to both endodermal and mesodermal lineages have not previously been described in the lung, nor indeed in any organ. In other studies, Chapman and colleagues [102] identifi ed a novel integrin α6β4-expressing alveolar epithelial cell in the mouse lung that can repair the lung following ALI. Hegab and colleagues [103] provide convin cing evidence that submucosal gland duct cells constitute a stem/progenitor cell population. Taken together, these studies off er considerable promise for a therapeutic role for endogenous lung stem/progenitor cells in lung diseases such as ALI/ARDS.

Barriers to clinical translation
It is not known whether the regenerative capacity demon strated by these c-kit-positive stem cells can be replicated in human lung, and indeed whether the tissue produced will function as normal lung. Likewise, the feasibility of harvesting endogenous stem cells from a critically ill ALI/ARDS patient is unclear. It is likely that the fi rst clinical applications of lung stem cells will be in the fi eld of tissue bioengineering for chronic conditions such as tracheal atresia

Conclusion
Stem cells constitute a promising therapeutic strategy for patients suff ering from ALI/ARDS. MSCs appear closest to clinical translation, given the evidence that they may favourably modulate the immune response to reduce lung injury, while maintaining host immune-competence and also facilitating lung regeneration and repair. Th e demonstration that human MSCs exert benefi t in the endotoxin-injured human lung is particularly persuasive. However, gaps remain in our knowledge regarding the mechanisms of action of MSCs, the optimal MSC administration and dosage regimens, and the safety of MSCs in critically ill patients. It is anticipated that these remaining knowledge defi cits will be addressed in ongoing and future studies. Other stem cells, such as ESCs and iPCs, are at an earlier stage in the translational process, but off er the hope of directly replacing injured lung tissue. Ultimately, lung-derived stem cells may off er the greatest hope for lung diseases, given their role in replacing and repairing the native damaged lung tissues.