Regenerative Pulmonary Medicine: Potential and Promise, Pitfalls and Challenges

Abstract

Lung disease is an increasing public health problem worldwide. According to the American Lung Association, more than 400,000 people die of lung diseases in the United States each year, which accounts for one in every six deaths overall. These staggering figures translate into a cost of more than $100 billion per year [1]. Even more concerning is the fact that in many chronic lung diseases, we have no therapeutic interventions with which to arrest or reverse the pathobiology of these destructive processes, or to restore functional lung tissue. Thus, we treat patients’ symptoms, but the underlying diseases continue to progress. In these circumstances, our therapeutic options ultimately turn to lung transplantation once diseases such as chronic obstructive pulmonary disease (COPD)/emphysema, idiopathic pulmonary fibrosis, cystic fibrosis, and idiopathic pulmonary arterial hypertension (PAH) become end-stage. Lung transplantation is a life-prolonging procedure for many patients; however, there is a shortage of available donor lungs, and, even when transplanted, the average survival for adult lung recipients is approximately 5–6 years [2]. Recipients are vulnerable to transplant-related diseases, such as bronchiolitis obliterans syndrome, which limits long-term survival in many patients [2],[3]. Thus, there is a desperate need for new and innovative therapies for a number of chronic lung diseases, including diseases that develop after lung transplantation.

The Use of Stem Cells and Regenerative Therapies in Pulmonary and Pulmonary Vascular Diseases: An Overview

Much of the evidence for the presence of lung progenitor cells comes from studies in mice in which selective ablation of epithelial cells resulted in epithelial regrowth. These types of studies, combined with genetic lineage tracing experiments, suggest the presence of facultative progenitor cells, or cells that exhibit a differentiated phenotype under basal conditions but are able to proliferate for tissue maintenance or in response to injury [4–6]. These reparative cells have been found in distinct anatomical regions of the lung, including the trachea/proximal airways, distal airways, and alveoli [7–20]. The challenge for the application of these cells therapeutically, however, is the inability to expand them ex vivo adequately to allow for exogenous administration in order to repair an injured lung in vivo. In addition, while in certain diseases repair of airway epithelial cells may be very advantageous, many of the chronic, destructive diseases leading to respiratory failure involve more than epithelial cell loss or dysfunction. Destructive processes alter the global architecture of the lung, leading to deterioration in function. At the core of the problem is disruption of the functional respiratory unit of the lung, the alveolar-capillary unit. Thus, not only is there need for repair of the epithelium within the proximal and distal airways, but regeneration of alveoli and their associated capillaries must also occur to maintain physiologic homeostasis and optimize lung function. Thus, for this type of cell therapy to move forward, an undifferentiated stem cell is needed with the ability to regenerate cells from more than one lineage, and to regenerate tissue originating from even more than one germ layer.

The cellular and anatomic complexity of the lung coupled with the fact that its largest functional volume fraction is composed of air space not only dictate a need for a true undifferentiated cell, but also require a novel method for delivery that will facilitate homing and persistence of intravenously delivered cells. Recent work suggests that intra-alveolar delivery of homing molecules in the form of nanoglycan particles may provide just the necessary homing signal, should the ideal cell be identified [21].

Since the lung is a complex organ, originating from cells of more than one germ layer, pluripotent cells would serve as an ideal means of treating lung diseases [22]. Embryonic stem (ES) cells, and, more recently, induced pluripotent stem (iPS) cells, are being investigated for their therapeutic potential. Embryonic stem cells are derived from preimplantation blastocysts [23],[24], and are capable of differentiating into cells of all three germ layers. Bioethical controversies exist, however, regarding the derivation and use of ES cells. In addition, ES cells in the undifferentiated state have been shown to form teratomas in vivo, for which reason attempts have been made to initiate the differentiation process prior to their administration. Unfortunately, differentiation of ES cells prior to administration may eliminate their potential to generate such a complex tissue as the lung. Owing to these and other issues, iPS cells with pluripotent properties have been considered an alternative to ES cells [25]. To generate these cells, retroviral transfection of somatic cells with specific transcription factors is used to induce a pluripotent phenotype; however, retroviral induction limits their potential use in humans. Ways to remove viral sequences are being investigated, and the use of non-viral means to transduce the cells are being explored [26]; yet, many issues need yet to be resolved [27], including the potential for insertional mutagenesis and tumor formation [28], before the use of these cells can be considered therapeutically. Nevertheless, this field is developing rapidly and great effort is being directed toward clinical application of iPS cells in the future.

We recently identified a stem cell from human lung tissue (human LSCs), using the stem cell antigen, c-kit, as a marker of identification and characterization [29] (vide infra). The LSCs were shown to be self-renewing, clonogenic (with generation of large clones), and multipotent both in vitro and in vivo. Moreover, from a single cell preparation, the cells could be expanded, administered to a mouse after lung injury, and result in the formation of human bronchioles, alveoli, and pulmonary vessels [29]. These human structures were fully integrated into the damaged mouse lung, both at the level of the airways and the vessels. Thus, we were able to generate lung tissue of both endodermal and mesodermal origins in the mouse with a single precursor cell. We believe that the identification of a human lung stem cell, with these regenerative properties, significantly advances the future potential for stem cell therapy in human lung disease.

The feasibility of using an endogenous c-kit^pos stem cell isolated from another organ in human disease was recently shown in the SCIPIO trial [30]. Autologous cells were isolated from the hearts of patients with ischemic cardiomyopathy, expanded ex vivo, and then returned to the patients from whom they were isolated at a later date. This work demonstrated the ability to isolate c-kit^pos stem cells from a diseased organ, to expand the cells effectively for therapeutic use, and to administer them safely to patients. Even with a limited number of patients in this phase I clinical trial, initial data showed a clear benefit in organ function. Taken together, this work provides encouraging support for the use of organ-derived c-kit^pos stem cells in human diseases in general.

To date, other pre-clinical studies using cells for therapeutic intervention in animal models of lung disease have focused on cells harvested from sources outside the lung. Many studies have utilized bone marrow-derived cells. Interestingly, investigators have found that a subpopulation of adherent cells (human or murine) from the bone marrow express Clara cell secretory protein (CCSP) [31]. When cultured in special conditions (air-liquid interface), these CCSP^pos cells begin to express other markers of epithelial cells, such as markers of type I and type II alveolar epithelial cells and of basal epithelial cells [31]. Bone marrow-derived CCSP^pos cells were able to home to sites of injury in a naphthalene-induced lung injury model, and there was evidence that these cells could repopulate the airway epithelium. In addition to these marrow-derived adherent cells, a myeloid-like progenitor cell harvested from the bone marrow has been reported to restore lung structure in neonatal mice exposed to hyperoxia, a model of bronchopulmonary dysplasia [32].

Other cells originating in the bone marrow that have been studied as potential therapeutic options for pulmonary diseases are mesenchymal stromal cells (MSCs) (vide infra), which have been the most widely studied. MSCs can be isolated from a number of tissues [33–37], including the bone marrow, and expanded ex vivo for application in animal models of lung disease. In preclinical animal studies, MSCs have been shown to contribute to tissue repair despite rare occurrences of engraftment and transdifferentiation, likely related to paracrine effects of the cells [38]. It appears that a critical property of MSCs is to limit tissue injury by modulating the immune response.

Mesenchymal stromal cells have been shown to be beneficial in a number of adult animal models, including acute lung injury, sepsis, allergic airways inflammation, bronchiolitis obliterans, elastase-induced emphysema, cigarette smoke-induced airway enlargement, pulmonary fibrosis induced by bleomycin, ischemia-reperfusion injury, and pulmonary hypertension (induced by monocrotaline or hypoxia), as reviewed recently [5],[39] and described in more detail below. Moreover, the potential benefit of MSCs is not limited to adults, as Aslam and colleagues have shown that MSCs attenuate lung injury in a neonatal model of murine bronchopulmonary dysplasia (BPD) [40]. Ongoing issues under consideration in MSC studies include not only the candidate diseases for future therapy, but also the appropriate route and timing of cell administration. Whether cells, conditioned medium, or specific secretory products derived from MSCs are adequate for their beneficial response is also currently under study [41]. Pre-clinical studies show therapeutic promise for MSCs, particularly for lung diseases with an ongoing inflammatory response driving disease pathobiology. A clinical trial of MSCs in patients with moderate to severe COPD (http://clinicaltrials.gov, Identifier: NCT00683722) was recently completed. The primary goal of the trial was to determine the safety of MSC infusions in patients with lung disease, and, secondarily, to determine whether MSCs could decrease chronic inflammation and improve lung function and quality of life in patients with COPD [5], [42]. The full results of the trial are soon to be forthcoming.

The other cell type presently undergoing clinical trials for lung disease is the endothelial progenitor cell (EPC). EPCs were originally described as a population of mononuclear cells in the blood capable of differentiating into endothelial cells in vitro [43]. It has recently become evident, however, that two distinct subsets of EPCs exist: an early outgrowth EPC derived from a hematopoietic lineage (initial cell described), and a late outgrowth EPC derived from an endothelial lineage [44], [45]. Studies are ongoing to understand the importance of circulating levels of EPCs in various lung diseases, and to explore the determinants of mobilization and recruitment of endogenous EPCs to the lungs [5], [39].

Owing to the complex architecture of the lung and the challenges of regenerating tissue in diseases (like emphysema) that lack a structured matrix upon which stem cells can reconstruct the lung, tissue engineering has also become an active area of investigation in pulmonary regenerative medicine. Tissue engineering refers to the generation of functional tissue that can replace endogenous tissue lost due to disease or injury [46]. For tissue engineering to be successful, however, a major challenge that needs to be overcome is the selection of the best cell source(s) to build both the lung parenchyma and its vascular supply. A successful approach would likely require either an undifferentiated, multipotent stem cell with the ability to generate numerous cell types, or a mixture of numerous unipotent, lineage committed cells. In addition, appropriate scaffolds or matrices need to be developed to support the three-dimensional regeneration of the tissue. In an organ like the lung, one also must consider whether a region of the parenchyma (such as the distal alveoli and vascular supply) or the proximal airways need reconstruction. While this distinction and its regenerative challenges are very onerous to contemplate, recent articles have demonstrated progress on both fronts. For large airways, investigators have shown that a bronchus could be bioengineered from a decellularized donor trachea, which was then colonized with epithelial cells and chondrocytes derived from MSCs [47]. This graft was then used to replace the left main bronchus of a recipient patient. The generation of more distal airways and alveolar structures is not as advanced as the generation of large airways, and is presently in pre-clinical studies. Cortiella and colleagues seeded somatic lung progenitor cells onto synthetic scaffolds (polyglycolic acid or pluronic F-127) in vitro, which resulted in structures similar to the distal lung [48]; however, when these cells were seeded on the synthetic scaffold and implanted in vivo, an inflammatory response ensued that was detrimental for lung tissue morphogenesis. In a recent advance, rat lungs were decellularized by detergent perfusion, leaving a scaffold of acellular airways, alveoli, and vessels. The scaffold was then seeded with epithelial and endothelial cells, and allowed to regenerate in a whole organ culture system [49]. The bioengineered lungs were then orthotopically transplanted into recipient rats, which had undergone pneumonectomy. These rats showed improved arterial blood gases compared with pneumonectomized control rats, and were able to be maintained without ventilatory support for six hours after transplantation. In a comparable set of experiments, Peterson and colleagues decellularized rat lungs, and then took the acellular lung matrix and mounted it inside a bioreactor that allowed seeding of vascular endothelial and airway epithelial cells [50]. After days of culture, the engineered lungs were removed from the bioreactor and implanted into syngeneic rat recipients for 45–120 minutes. Impressively, these engineered lungs participated in gas exchange following implantation. Taken together, these studies show the potential applications of tissue engineering in the future of pulmonary regenerative medicine.

To advance the fields of stem cell therapy and tissue engineering in diseases of the lung requires that a number of challenges be overcome. These include the type(s) and numbers of cells to be used, the best timing and route of their administration, and, for tissue engineering purposes, the best scaffold for the construction of a three dimensional matrix. We believe that the primary issue in successful stem cell therapy/tissue engineering is the selection of the optimal cell(s) for therapy. The discovery of an undifferentiated, multipotent stem cell in the human lung, which can be adequately expanded ex vivo for use in pre-clinical trials [29], provides great advantages in an organ like the lung, which is made up of over 40 distinct cell types and is structurally complex. We will next discuss in more detail the use of EPCs, MSCs, and, lastly, these newly discovered endogenous lung stem/progenitor cells.

Endothelial Progenitor Cells (EPCs)

The use of ex vivo expanded EPCs derived from blood or bone marrow as a potential therapy for lung disease has predominantly focused on pulmonary hypertension. In a model of monocrotaline (MCT)-induced pulmonary hypertension, Zhao and colleagues demonstrated that early outgrowth EPC could prevent pulmonary hypertension when given early after MCT administration [51]; however, when administered late after the onset of marked elevations in right ventricular pressure, EPCs were only partially effective. The investigators used gene transfer techniques to enhance endothelial nitric oxide synthase (eNOS) expression in EPCs, and then tested these cells in the MCT-induced model. They found that eNOS gene transfer enhanced the ability of EPCs to restore pulmonary hemodynamics and vascular architecture in this model [51]. Owing to these pre-clinical results, an early-phase clinical trial was initiated (http://clinicaltrials.gov, Identifier: NCT00469027) to assess the tolerance of eNOS-enhanced autologous EPCs delivered to the pulmonary circulation of patients with severe PAH refractory to conventional treatment [52]. The Pulmonary Hypertension And Cell Therapy (PHACeT) trial, a dose-escalation study, is also ongoing at the University of Toronto using these cells. Results from two other small clinical trials assessing intravenous infusion of autologous EPCs in patients with idiopathic PAH, in conjunction with conventional therapy, suggest that EPCs are safe and potentially beneficial [53], [54].

While PAH is a disease process driven by pathology of the pulmonary vasculature, right ventricular failure ultimately leads to death in these patients. Pre-clinical data in rats suggest MSCs improve right ventricular function in MCT-induced pulmonary hypertension [55]. In this animal model, the improvement in right ventricular function may relate to the improvement in pulmonary vascular pathology; however, this does raise the question of whether a cardiac stem cell may also be of benefit, not only in diseases like ischemic cardiomyopathy [30], but also in idiopathic PAH when given in conjunction with conventional therapy to improve right ventricular function and prevent right heart failure.

Mesenchymal Stromal Cells (MSCs)

The existence of bone marrow stromal cells distinct from hematopoietic progenitor cells was first suggested in the 19^th century by Cohnheim, a German pathologist investigating experimental models of wound repair (reviewed in [34]). A century later, Friedenstein characterized plastic adherent cells isolated from the bone marrow (BM) that were distinguishable from the majority of hematopoietic cells by their fibroblast-like appearance and the ability to form cells of the osteogenic, chondrogenic, and adipogenic lineages [56]. Several reports have since demonstrated that systemically-injected mouse bone marrow-derived stem cells differentiate into parenchymal cells of various non-hematopoietic tissues, including muscle, cartilage, bone, liver, heart, brain, intestine, and lung [34], [57–59], and are able to adopt the morphological and molecular phenotype of their new resident organ. The nomenclature of marrow stromal cells has evolved over the years to mesenchymal stem cells and, most recently, to multipotent mesenchymal stromal cells or MSCs [60]. In addition to bone marrow, MSCs have been isolated from multiple tissues, including umbilical cord blood, Wharton’s jelly, placenta, and adipose tissue [61–65]; shown to have immunomodulatory properties; and demonstrated significant therapeutic potential in several animal models of human disease. Because investigators used different methods of isolation and expansion and developed different approaches for characterizing these cells, the Mesenchymal and Tissue Stem Cell Committee of the International Society for Cellular Therapy proposed minimal criteria to define human MSCs [60]. These included (i) plastic-adherence; (ii) cell surface marker expression of CD105, CD73, and CD90, and lack of expression of CD45, CD34, CD14 or CD11b, CD79alpha or CD19, and HLA-DR surface molecules; and (iii) differentiation into osteoblasts, adipocytes, and chondroblasts in vitro. While these criteria will evolve as new knowledge unfolds, this minimal set of standard properties will foster a more uniform characterization of MSCs and facilitate the exchange of data among investigators.

Despite the controversy surrounding the true characteristics of MSCs, their therapeutic potential has been harnessed to treat a wide variety of diseases in the laboratory, and is now being investigated in patients [66]. An increasing number of recent studies report the therapeutic potential of exogenously administered MSCs in experimental models of lung injury (reviewed in [67–69]); however, whether exogenous or lung resident MSCs can participate in processes of cell renewal and repair remains inconclusive. MSCs have been isolated from tracheal aspirates of ventilated human preterm infants and shown to express MSC markers; were negative for hematopoietic or endothelial cell markers; and correlated with prolonged ventilator dependence, oxygen requirement, and the development of BPD [70], [71]. Lama and colleagues reported successful isolation of MSCs from the bronchoalveolar lavage of adult lung transplant patients [72] and showed that in seven sex-mismatched transplant recipients, the MSCs were donor-derived up to 11 years following transplantation, suggesting that connective tissue cell progenitors reside in the lung.

The potential use of MSCs for the treatment of a multitude of diseases, including diseases of the lung, is based on their low immunogenicity and ability to evade clearance by the recipient host immune system. This phenotype is partly achieved through their low expression of the major histocompatibility complex I and II proteins and their ability to inhibit proliferation and function of immune cells, including T cells, B cells, natural killer cells, and dendritic cells (reviewed in [5]). In addition, MSCs can be isolated from readily available tissues, including the umbilical cord and bone marrow, and expanded ex vivo, as well as be genetically-manipulated to deliver cytoprotective genes of interest to diseased organs. These properties, combined with their potent immunomodulatory effects, make MSCs attractive candidates for treating inflammatory diseases. Since inflammation underlies many lung, cardiovascular, and multisystem disorders, MSC therapy can have therapeutic implications for a broad number of disorders.

In the adult lung, most studies to date have focused on the proof-of-principle that MSCs can home to the lung and adopt various lung cell phenotypes in different injury models. Although these data are at times conflicting, in all of these models of disease, exogenous administration of MSCs showed therapeutic benefit in preventing lung injury (reviewed in [67–69]). Recently, the therapeutic potential of MSCs was also reported for the developing lung using neonatal mouse and rat models of BPD [40], [73], [74].

In 2001, Krause and colleagues, reported on the multipotent nature of bone marrow-derived cells, and opened the door to the potential use of adult stem cell therapy for tissue replacement [59]. Expanding on previous work, they demonstrated that upon administration of bone marrow CD34^pos lin^neg cells from male donor mice into irradiated female recipients, up to 20% of alveolar epithelial cells (AECs) were derived from the donor animal and persisted up to 11 months following transplantation [59]. Consistent with this first observation, other investigators reported that various preparations of bone marrow-derived cells were able to engraft and adopt the phenotype of distal lung cells, including type I [75] and type 2 AECs [76] as well as lung fibroblasts [77], in a variety of experimental models. These studies demonstrate the potential of bone marrow-derived cells to home to the injured lung, engraft, and adopt the phenotype of one or many of the lung cells. In addition, prior injury amplifies stem cell engraftment, as observed in other tissues [78].

Conversely, more recent studies do not support adult stem cell plasticity in the lung. Engraftment was assessed in chimeric animals [bone marrow ablation and reconstitution with GFP-labeled hematopoietic stem cells (HSCs)] and in parabiotic animals (created by joining the circulation of a transgenic GFP mouse with the circulation of a wild type mouse) [79]. While the bone marrow-ablated group showed robust reconstitution of circulating HSCs and the parabiotic animals had robust hematopoietic chimerism, there was little evidence for stem cell engraftment and transdifferentiation in the brain and the liver, and none detected in the lung [79]. Two other studies investigated lungs of recipient mice transplanted with bone marrow from transgenic mice that ubiquitously express eGFP or LacZ: engrafted cells displayed type II AEC characteristics [80]. Deconvolution microscopy, however, in which a three-dimensional image of the lung was created, determined that the co-expression of surfactant protein-C (SP-C) and eGFP was a false positive signal: the SP-C signal resided just outside the eGFP^pos cells [80]. This example highlights the need for precise tools with which to assess stem cell engraftment, including confocal microscopy [81]. Mice whose bone marrow was reconstituted with bone marrow-derived cells from SP-C-eGFP donor mice showed no evidence of donor cells becoming type II AECs using three antibody independent assays, fluorescence-activated cell sorting (FACS), fluorescence microscopy, and real time PCR [82]. In humans, bone marrow-MSC repopulation also seems to contribute minimally to the type II AECs after cross-sex lung transplantation [83].

Despite the controversy about stem cell engraftment, evidence suggests that MSCs have the capacity to home to the lung, and, importantly, afford significant lung protective effects likely through a combination of mechanisms: these may include transdifferentiation of a small subset of cells, direct cell-to-cell connections with transfer of cell components, release of soluble factors, as well as release of other cellular components, such as microvesicles or exosomes. Depending on the disease context or stage of lung development, MSCs and their secreted soluble factors and cellular components lead to lung growth and repair. These effects are likely a consequence of a combination of immune modulation; alteration of vascular and epithelial cell responses; and stimulation of endogenous lung stem cells to proliferate, differentiate, and participate in lung repair, orchestrated by paracrine MSC-derived mediators. Evidence for these MSC-dependent functions will be summarized below in studies using experimental models of lung disease.

In mouse models of acute lipopolysaccharide (LPS)-induced lung inflammation, fibrosis, and emphysema, MSCs appear to abrogate lung injury. Intratracheal LPS injection caused rapid mobilization of bone marrow cells into the circulation [84]. Bone marrow cells accumulated within the inflammatory site and differentiate to become endothelial and epithelial cells. The suppression of bone marrow cells by sublethal irradiation before intrapulmonary LPS led to emphysema-like changes and reconstitution of the bone marrow prevents these changes [84]. Likewise, exogenous intratracheal administration of MSCs improved survival, attenuated lung inflammation, and limited lung permeability [85]. MSCs transfected with the vasculoprotective gene, angiopoietin-1, also appeared to protect against inflammation and increased lung permeability [86].

Intravenously administered bone marrow-MSCs engraft to areas of fibrosis in models of bleomycin-induced lung injury and exhibit an epithelium-like morphology [87]. MSC administration immediately following bleomycin attenuated lung inflammation and fibrosis [87]. In the same model, myelosuppression increased the susceptibility to injury, and MSC administration improves survival and attenuates fibrosis [88].

Adipose tissue-derived stromal cells induce alveolar regeneration and improve lung function in elastase-induced emphysema in rats [89]. Likewise, sex-mismatched bone marrow-MSC transplantation in irradiated female rats improved lung architecture, and Y chromosome fluorescence in situ hybridization (FISH) and immunohistochemical staining for SP-C confirmed that MSCs engrafted in recipient lungs and differentiated into type II AECs [90]. Intratracheal MSC delivery 14 days after elastase injury decreased lung injury and IL-1β levels, and resulted in transient increases of lung epidermal growth factor, hepatocyte growth factor, and secretory leukocyte protein inhibitor through postulated paracrine mechanisms as there was infrequent MSC engraftment and differentiation detected [91].

Using the rat MCT model of pulmonary hypertension, intratracheal MSC administration improved cardiopulmonary physiology [92], and treatment with MSCs overexpressing eNOS resulted in more pronounced improvement of PH-associated right ventricular (RV) hypertrophy with better survival compared to treatment with unmodified MSCs [93]. Similarly, transplantation of prostacyclin synthase-transduced MSCs reduced pulmonary pressures and improved survival in the same rat MCT model of pulmonary hypertension [94].

We have previously reported that heme oxygenase-1 (HO-1)-deficient mice develop severe pulmonary hypertension with right ventricular infarction and eventual failure under prolonged hypoxia [95], [96]. Transplantation with MSCs derived from transgenic mice harboring a human HO-1 transgene under the control of SP-C promoter reversed established disease in wild type recipients, and, importantly, prevented RV infarction and death in HO-1 null recipients even when delivered after five weeks of hypoxia [97]. We observed that intravenous MSC delivery inhibited the induction of pro-inflammatory cytokines, monocyte chemoattractant protein-1 (MCP-1), and interleukin-6 (IL-6), but increased levels of the anti-inflammatory cytokine, IL-10, and prevented the alveolar macrophage accumulation that is a response to early hypoxia exposure [97]. Although robust protection was observed against pulmonary hypertension, only a small fraction of cells were retained in the recipient lung, as has been true for most of the other animal models of lung injury.

Intratracheal delivery of bone marrow-derived MSCs in newborn rats exposed to hyperoxia improved survival and exercise tolerance, while attenuating alveolar and lung vascular injury in this model of neonatal lung injury that recapitulates the findings of human BPD [73]. Chang and colleagues [74] reported that intratracheal delivery of umbilical cord-derived human MSCs was more effective in attenuating hyperoxia-induced lung injury in the neonatal rat model compared with intraperitoneal transplantation. Our group similarly reported significant improvement in neonatal murine hyperoxic lung injury with intravenous delivery of mouse bone marrow-derived MSCs; however, we observed complete inhibition of lung disease and pulmonary hypertension when neonatal hyperoxic mice were treated with MSC-conditioned media resulting in preservation of normal lung architecture and, thus, demonstrating for the first time, the in vivo efficacy of cell-free media supplanting cell therapy [40].

Taken together, these observations suggest that an intact bone marrow serves to limit the extent of lung injury, that exogenously administered MSCs home preferentially to sites of injury, and that these MSCs can adopt distal lung cell phenotypes and prevent lung injury. In addition, MSCs may also be used as vehicles to target gene delivery in vivo [86], [93], [97] in ischemia-reperfusion associated with lung transplantation targeting IL-10 delivery [98], and for the cystic fibrosis transmembrane conductance regulator (CFTR) gene with successful, epithelial cell-specific expression [99]. Wang and colleagues reported that MSCs in co-culture with AECs from ΔF508 cystic fibrosis patients acquire expression of the wild type CFTR protein, adopt a columnar or epithelial shape of the human airway epithelium, and express cytokeratin-18 [100]. Cardiomyocytes and human MSCs were shown to communicate in co-culture through small diameter nanotubes allowing transfer of mitochondria from MSCs to cardiomyocytes [101] and, in another report, to rescue ischemic cardiomyocytes from cell death [102]. In all of these studies, however, the low numbers of engrafted MSCs seems insufficient to account for the therapeutic response attributed to these consequences of direct cell-cell contacts.

Based on these and other reports, increasing evidence has accumulated to indicate that soluble factors secreted by MSCs account for many of their observed beneficial effects. In vitro, cells from bleomycin-injured, but not from normal, mouse lung produce soluble factors that cause MSC proliferation and migration toward the injured lung [88]. Furthermore, conditioned medium obtained from MSCs blocks the proliferation of an IL-1α-dependent T cell line and inhibits production of tumor necrosis factor-α by activated macrophages in vitro [103]. In endotoxin-induced systemic inflammation, BM-derived MSCs decrease both the systemic and local inflammatory responses [104]. These effects did not require either lung engraftment or stem cell transdifferentiation, suggesting that humoral and physical interactions between stem cells and lung cells account for some of the beneficial effects. Furthermore, in further support of paracrine mechanisms acting in vivo, administration of cell-free conditioned medium from MSC was more effective than MSCs themselves in inhibiting lung disease and preserving alveolar growth in the neonatal mouse model of BPD [40].

The list of candidate mediators released or induced by MSCs is expansive and includes IL-I receptor antagonist, prostaglandin E₂, keratinocyte growth factor, angiopoietin-1, indoleamine 2, 3-dioxygenase, transforming growth factor-β, tumor necrosis factor-α-stimulated gene/protein 6, and IL-10, among others (reviewed in [105]). No single molecule is likely to mediate the broad cytoprotective and long-lasting actions of MSCs, however, and it is the ability of MSC-derived paracrine mediators, perhaps packaged and secreted as membrane microvesicles containing proteins and/or genetic information (mRNAs, microRNAs) capable of modulating multiple signaling pathways, that could result in therapeutic efficacy. Additionally, MSCs may stimulate organ-specific progenitor stem cells to expand, proliferate, and differentiate, and to participate in cellular repair to replace injured cells. The paracrine mechanisms of MSC action represent an active area of investigation necessary for the eventual successful application of stem cell-based therapies for inflammatory diseases, including those of the lung.

Endogenous Lung Stem/Progenitor Cells

The possibility that the human lung is a self-renewing organ characterized by a compartment of endogenous stem/progenitor cells that control the growth and turnover of epithelial and vascular cells imposes a reevaluation of the biological mechanisms implicated in the maturation of the lung after birth, the maintenance of organ homeostasis in the adult, and tissue repair after injury. The transition from the fetal to the adult respiratory system involves abrupt changes in pulmonary volume. Lung growth continues for several years after birth and is characterized by a ~20-fold increase in alveolar number [106], [107], pointing to alveolar regeneration as a major determinant of the acquisition of the adult organ phenotype in humans. The expansion of the alveolar compartment [108] has to be coupled with a comparable adaptation of the pulmonary capillaries, which are needed to form functionally integrated alveolar-capillary units. The growth of the capillary microcirculation may be promoted by recruitment of circulating vascular progenitors, angiogenesis with sprouting of existing vessels, or vasculogenesis mediated by activation and differentiation of endogenous stem cells [109], [110].

Similar regenerative processes may regulate the preservation of the steady-state of the adult lung, and the cellular adaptations following tissue damage. Whether dividing epithelial and vascular cells are transit-amplifying cells generated by commitment of endogenous or exogenous stem cells, or constitute a pool of differentiated cells that reenter the cell cycle and replicate, remains controversial. Additionally, epithelial cells may retain a certain degree of developmental plasticity, dedifferentiate, and acquire a proliferative phenotype, generating cells with specialized function. Understanding the origin of the pulmonary structures and the mechanisms that modulate lung homeostasis offers the opportunity to enhance this naturally occurring process, promoting tissue restoration following injury.

Several stem cell-regulated organs were recognized to adhere to a hierarchical model of cell growth; however, this view has been challenged in the hematopoietic system, skin, intestine [111], and, lately, in the lung [112]. The airway epithelia comprise the tracheobronchial, bronchiolar, and capillary-alveolar units. Each of these regions is composed of specialized epithelial cells, which, together, account for more than 40 phenotypically and functionally different pulmonary cell types. The complex architectural organization of the lung has prompted the hypothesis that pulmonary cells do not conform to the classic stem cell hierarchy [112]. This theory is supported by the identification of several independent putative stem cell pools, which reside in anatomically distinct domains and form segment-specific epithelial cells, sustaining regional homeostasis and repair [113]. These partially committed progenitors scattered throughout the airway epithelia, however, do not fulfill the functional definition of stem cells since they are not multipotent [114]. The recently identified human lung stem cell fulfills the criteria of a bona fide stem cell which differentiates into specialized cells of endodermal and mesodermal lineages and reconstitutes normally appearing epithelial and vascular structures in the injured mouse lung [29]. These observations support the notion that pulmonary cells are organized according to a hierarchical model of organ growth, homeostasis, and repair [115].

In the adult mouse lung, several approaches have been implemented for the detection of resident progenitor cells: they include the recognition of putative niches by pulse-chase label assays, the use of fate mapping strategies, and the induction of survival and proliferative responses to injury [116]. The milieu of the niche is crucial for the maintenance of the lineage negative phenotype of stem cells, which, outside of the niche, inevitably undergo commitment and differentiation [117]. The search for organ-specific niches involves the identification of cells with the phenotypical properties of stem cells, the demonstration that stem cells are physically connected to supporting cells, and the documentation that a lineage relationship exists between ancestors and descendents [117]; however, the evidence of interstitial, intraepithelial, or intravascular structures with the architectural organization of niches is lacking in the mouse lung. The traditional pulse-chase assay is a well-established protocol for detecting stem cells and their niches in adult organs [116]. It consists of the administration of thymidine analogs for a short interval, which is followed by a long chase period to allow the dilution of the label in frequently dividing cells. At the end of the chase, the persistence of a strong signal intensity characterizes rarely replicating cells that retain the halogenated nucleotide incorporated during the pulse period. Although, the quiescent property of stem cells makes this protocol suitable for their recognition, corroborative data are required for the unequivocal documentation that long-term label-retaining cells (LT-LRCs) are true stem cells and that the surrounding environment is an actual stem cell niche.

The presence of LT-LRCs in the submucosal gland (SMG) of the proximal airways, and the observation that repair of the tracheal epithelium following naphthalene injury is accompanied by migration of cells hosted in the SMG, have suggested that these discrete structures contain stem cells [118], [119]. Paradoxically, the identity and fate of these alleged progenitor cells are at present unknown. The intercartilagenous zone of the mouse trachea has been considered a niche for basal epithelial cells (BECs), which have been viewed as a lung stem cell class [14], [120], [121]. In mice and humans, BECs are positive for the transcription factor p63 and the intermediate filament cytokeratin 5 [14]. The uniform expression of p63 and cytokeratin 5 in the basal epithelium of human distal airways raises concerns about the appropriateness of these two markers for the identification of stem cells, which constitute a rare cell population in any given organ. Although the presence of p63 characterizes basal cells during development [122], the formation of the trachea is not prevented in p63 knockout mice [123–125]. The tracheal epithelium lacks BECs, but shows both ciliated and mucosecretory cells, questioning the progenitor nature of BECs during embryonic-fetal life and early after birth [125]. In a manner similar to the lung, p63 is expressed in the basal layer of the epidermis but is absent in skin stem cells located in the ‘bulge’ [126]. Individual p63-negative bulge stem cells form holoclones in vitro, and the clonal progeny derived from a single founder cell reconstitute in vivo hair follicles, epidermis, and sebaceous glands [126]. In contrast, BECs are not clonogenic and their epithelial specification is consistent with the ability to differentiate exclusively into tracheal epithelial cells: they do not generate type 1 or type 2 AECs or pulmonary vessels [114].

In mice developing a severe form of acute respiratory distress syndrome following infection with a sublethal dose of H1N1 influenza virus, a p63-positive population of cells has been detected in the bronchiolar epithelium [20]. These cells, which are absent in healthy mice, form structures that resemble alveoli, raising the possibility that p63-expressing cells may participate in alveolar reconstitution after specific injuries but do not contribute to the homeostasis of distal airways [20]. In a recent perspective [127], questions have been raised concerning the origin of these cells, which are typically restricted to the tracheal lining. Three possibilities have been advanced: transient reseeding of BECs from the proximal airways following damage, dedifferentiation of a small subset of Clara cells, or derivation from more immature cells, including the c-kit^pos LSCs identified in the human lung (Figure 1). In all cases, BECs do not appear to constitute resident bronchiolar stem cells.

Figure 1. Localization of hLSCs in Bronchioles.

Bronchiole, ~1.2 mm in diameter, with epithelial cells, positive for pan-cytokeratin (pan-CK), and smooth muscle cells, positive for α-SMA (α-smooth muscle actin). Cartilage is also seen (bright blue). The basal cells of the epithelial lining express p63 in their nuclei (white). Several c-kit^pos cells (green) are present within the bronchiolar wall and its proximity. The 4 areas included in the rectangles are shown sequentially at higher magnification in the subsequent panels. In all cases, the basal epithelium also contains c-kit^pos cells which express p63 in their nuclei (white) and cytokeratin 5 (CK5, magenta; arrows) in their cytoplasm (see insets). The basal epithelium contains also c-kit^pos cells negative for p63 and CK5 (asterisks).

Clara cells have been claimed to represent the resident stem/progenitor cell pool of the distal airways [9]. Clara cells are distributed in the tracheal epithelium and partly in the bronchioles of the mouse and in the small airways of the human lung. These cells can divide, contributing to the repair of proximal and more distal respiratory structures [9], [128]; these properties are, however, insufficient to justify their inclusion as stem cells. Additionally, Clara cells secrete mucin that per se argues against the undifferentiated phenotype that defines stemness. Following naphthalene injury, Clara cells are largely depleted and the remaining viable LT-LRCs have been defined as variant Clara cells (Clara^v). These cells, which are positive for Sca-1 and the Clara cell secretory protein (CCSP), can efflux Hoechst dye, possibly reflecting a side-population progenitor category [129]. A partially overlapping class of pancytokeratin-positive side population cells has been found, but their behavior in vitro mimics MSCs [130].

In the distal airways of the mouse lung, the putative niche comprising the bronchioalveolar duct junction contains a pool of putative progenitors, the bronchioalveolar stem cells (BASCs) [12], [131]. These naphthalene-resistant cells express the epithelial cell antigen, CCSP, and the alveolar marker SP-C, documenting their epithelial specification. Consistent with their early committed state, BASCs form very small colonies in vitro, indicating that the founder cell is not a lineage-negative clonogenic stem cell, but represents a cell that has initiated the terminal differentiation program and is at the end of its proliferative lifespan. By analogy with hematopoietic stem cells [131], Bmi1 appears to be indispensable for the proliferation of BASCs, a property that has been documented in vitro and in vivo [131]. The participation of BASCs in the homeostasis and regeneration of alveoli has been challenged [13]. By lineage tracing of CC10-positive Clara cells, BASCs have been shown to repair the terminal bronchiolar epithelium but not the alveolar epithelium [13]. In addition to BASCs, a subpopulation of type II AECs has been proposed as stem/progenitor cells of the alveolar structures [132]. Type II AECs divide and form not only type II pneumocytes but can also form type I pneumocytes. Based on this growth potential, type II cells have been considered progenitors of the alveolar epithelium [132].

The expression of the c-kit receptor was employed as marker for the identification and characterization of bona fide stem cells in the human lung [29]. Human LSCs are negative for pulmonary lineage markers and do not express epitopes of HSCs or MSCs, excluding their potential derivation from the bone marrow. Lin ^neg c-kit ^pos cells possess the fundamental properties of stem cells -self-renewal, clonogenicity, and multipotency - in vitro and in vivo. The delivery of clonal human LSCs to cryoinjured mice results in the formation of human bronchioles, alveoli, and pulmonary vessels (Figure 2), providing strong evidence in favor of the crucial role that human LSCs may have in lung homeostasis and tissue regeneration following damage. While clonal expansion in vitro may alter the molecular phenotype of these cells, our ability to passage this clone from one mouse to another with virtually identical stem-like properties strongly supports the view that this cell is a true endogenous lung stem cell. Taken together, these results support the view that a classical stem cell hierarchy exists in the lung [115].

Figure 2. Regeneration on Injured Mouse Lung by hLSCs.

Panels A–C: The epithelial cells of newly formed alveoli are EGFP-positive (A) and pan-CK-positive (B). Panel C: merge. The area included in the rectangle is shown at higher magnification in the lower panels. EGFP-positive pan-CK-positive alveolar epithelial cells carry only human X-chromosomes (hX-Chr, white dots).

In addition to their therapeutic implications, these observations challenge the generally accepted belief that the lung is an organ lacking a hierarchical organization regulated by a compartment of resident stem cells nested in niches. Early committed clonogenic human LSCs express p63 and cytokeratin-5, CC10 and pan-cytokeratin, CC10 and SP-C, TTF1 and SP-C or aquaporin-5. These findings suggest that BECs, Clara cells, BASCs, and AECs may derive from lineage specification of c-kit ^pos human LSCs.

Collectively, studies in the mouse lung have argued that cells with the properties commonly attributed to stem cells have been identified. This conclusion has been based predominantly on the ability of these cells to enter the cell cycle, divide, and contribute to pulmonary repair following extensive airway injury [118], [119], [14], [120], [121], [20],[9], [128], [130], [12]. BECs, BASCs, Clara cells, SP cells, and type II AECs have well-defined specialized functions. Unlike progenitor/stems cells, they are not multipotent, and can only differentiate into epithelial cell types restricted to specific segments of the lung. Importantly, they do not create pulmonary vessels, an essential component for the generation of functional alveolar-capillary units. Owing to their proliferative capacity, these cells can, thus, be viewed as late precursors or transit-amplifying cells that may be important contributors to lung repair.

Fate mapping strategies, based on fluorescent reporter genes, are commonly used to track the origin of cells and their destiny in animals in which genetic manipulations are easily introduced [133], [134]. This approach would represent the ideal retrospective assay for the detection of regenerated lung cells since the expression of the fluorescent label can be placed under the control of promoters of genes coding for epithelial and vascular proteins. However, this protocol cannot be implemented in humans and, most importantly, it provides information at the level of cell populations that share the reporter gene, but does not demonstrate the self-renewal and multipotentiality of single stem cells in vivo. In fact, it is impossible to discriminate whether individual stem cells divide asymmetrically, i.e., self-renew, and whether the cell types of the tagged progeny derive from activation of one or several resident stem cells, i.e., manifest unipotency or multipotency.

With the exception of the c-kit^pos human LSCs [29], the various pools of putative progenitors reported thus far appear only after injury, questioning their role in routine tissue homeostasis. Experimentally, most of current understanding of the progenitor/progeny relationship and stem cell phenotype in the adult lung originates from research using lung injury models in the mouse [14], [121], [20], [9], [129], [13]. This limitation is enhanced by the striking difference between the cytoarchitecture of airways in mice and humans [115]: in small rodents, Clara cells are scattered throughout the entire airway, while, in humans, goblet cells are the predominant secretory cell type in the proximal airways and Clara cells are restricted to bronchioles. Additionally, BECs are present in the distal airways of the human lung and are located in the tracheal region of the mouse lung [14]. Importantly, isolation, amplification, and delivery of these cells to damaged lungs has not been performed, making it impossible to define the potential clinical import of the various classes of putative primitive cells. The recent identification of c-kit^pos human lung cells [29] that can be expanded ex vivo and delivered in vivo with the capacity to both create and repair lung structures of all lineages fulfills the criteria of stem cells and further adds to the important body of work on lung stem cell biology in development and repair.

Conclusions

Cell-based therapies for lung diseases hold tremendous promise. Over the last decade, numerous studies have demonstrated the ability of EPCs, MSCs, and most recently, LSCs to facilitate regeneration of lung tissue in a variety of models of pulmonary and pulmonary vascular diseases. Human studies are currently ongoing or recently completed for the earliest clinical trials of various progenitor cell types in lung diseases. While no definitive results have yet to demonstrate the benefits of these therapies, preclinical studies suggest that the next decade offers great promise for the future of regenerative lung therapies for human chronic pulmonary diseases.