Coming to terms with tissue engineering and regenerative medicine in the lung

Abstract

Lung diseases such as emphysema, interstitial fibrosis, and pulmonary vascular diseases cause significant morbidity and mortality, but despite substantial mechanistic understanding, clinical management options for them are limited, with lung transplantation being implemented at end stages. However, limited donor lung availability, graft rejection, and long-term problems after transplantation are major hurdles to lung transplantation being a panacea. Bioengineering the lung is an exciting and emerging solution that has the ultimate aim of generating lung tissues and organs for transplantation. In this article we capture and review the current state of the art in lung bioengineering, from the multimodal approaches, to creating anatomically appropriate lung scaffolds that can be recellularized to eventually yield functioning, transplant-ready lungs. Strategies for decellularizing mammalian lungs to create scaffolds with native extracellular matrix components vs. de novo generation of scaffolds using biocompatible materials are discussed. Strengths vs. limitations of recellularization using different cell types of various pluripotency such as embryonic, mesenchymal, and induced pluripotent stem cells are highlighted. Current hurdles to guide future research toward achieving the clinical goal of transplantation of a bioengineered lung are discussed.

lung diseases such as emphysema, chronic obstructive pulmonary disease (COPD), cystic fibrosis, interstitial fibrosis, and pulmonary hypertension cause significant morbidity and mortality worldwide (5). With greater understanding of the mechanisms that underlie these diseases, and development of novel therapeutic approaches, clinical management of these diseases has improved substantially. However, current therapies are not curative and largely rely on preventing exacerbations, blunting symptoms, and slowing progression of pulmonary fibrosis and pulmonary hypertension in particular. Lung disease management relies on the use of multimodal drug therapies, pulmonary and physical rehabilitation, and nontransplant surgical interventions (e.g., lung volume reduction surgery, removal of bullae, placement of stents etc.). Regardless of these advances, the current curative approach for end-stage disease is lung transplantation.

Despite improvements in perioperative management for lung transplantation, a number of limiting issues make the creation of effective alternatives to lung transplantation an urgent, unmet need. By far, the most critical issue is the acute shortage of transplantable donor lungs with unacceptably long wait lists. For example, more than 2,500 patients in the United States alone are on the wait list, with an average wait time of >1 yr (208), and an unfortunately high rate of mortality during the wait period (approximately 15–20 per 100 waitlisted), and deterioration of the recipient's health status in the interim. A separate critical issue is the condition and transplantable status of donor lungs that often suffer edema, atelectasis, or even pneumonia depending on the donor's comorbidities and approaches to preoperative lung handling. Here, long ischemic times that cause lung damage correlate directly with primary graft failure (63). Added to these perioperative issues are reduced outcomes and limited longevity following transplantation due to long-term organ and systemic effects of immune suppression, infections, and problems with graft rejections. Accordingly, effective alternatives to traditional lung transplantation are needed. In addition to whole-lung transplantations, an unmet clinical need exists for strategies to replace large extrapulmonary airways for indications such as tracheal stenosis and tracheobronchial tumors (216). From the vascular perspective, besides artificial grafts, native donor vessels to replace larger pulmonary vessels are also needed (100). Beyond these larger vessels, an obvious need exists to ensure that the structure and function of the blood supply of a transplanted lung matches the functional needs of the organ in terms of ventilation, prevention of edema, and clearance of toxins.

Issues in Lung Bioengineering

There is now a large research and clinical thrust toward engineering of whole organs or their component tissues with a long-term goal of transplanting them into humans. The field of tissue engineering is geared toward developing clinically viable approaches to recreating or enhancing human organs and tissues that recapitulate both the structure and function of the native organ. For example, engineering of replacement skin tissues is now well recognized (68, 187), and there has been substantial progress in other organ systems such as cartilage (45, 54, 107, 159) and bladder (53, 104, 138, 185, 212). With regard to the lung, technologies that mimic lung function such as cardiopulmonary bypass and extracorporeal membrane oxygenation are well developed and routinely used in in-hospital settings, but they suffer from being of short-term, nonambulatory use. Here, despite substantial improvements in the technologies needed for extracorporeal oxygenation and removal of CO₂, issues such as coagulation, immunogenicity, and proinflammatory responses limit the long-term use of these technologies. Furthermore, the need for pump-based devices and associated equipment to maintain flow will likely limit miniaturization. In this regard, pumpless technologies with low-resistance diffusion membranes such as the Novalung interventional lung assist device (Novalung, Heilbronn, Germany) are now being explored for use in intensive care settings for removal of CO₂ in patients experiencing trauma, infection, and acute respiratory disease syndrome or even as a bridge to lung transplantation (13, 15, 26, 46, 51). However, even though these emerging approaches hold substantial potential for bridging toward transplantation, unlike the increasingly popular ventricular-assist devices that are emerging therapies for treating heart failure, transportable lung-assist devices are still in the research phase (74–76, 188). Thus lung bioengineering is still a nascent field. The questions then become: what is the current status of bioengineered lung tissues, and what are the limitations to bioengineering the lung toward transplantation?

The challenge of cellular heterogeneity.

The native human lung consists of >40 cell types that define its overall structure and functions of maintaining (and appropriately modulating) oxygenation and ventilation; limiting environmental, allergenic, and infectious insults; clearing microbes and toxins; and preventing edema. A progressively branching airway network ends in 200-μm-diameter alveolar sacs that total 70–100 m² surface area for gas exchange. The airway network intersects a branching pulmonary vascular network to process the entire cardiac output of 5 l/min into capillaries that are separated from alveoli by a thin respiratory membrane. This incredible surface area-to-volume ratio on the one hand is remarkable in terms of gas exchange efficiency, yet it represents a major limitation in the ability to engineer biocompatible materials and recellularization on a large scale that would allow such efficient exchange of oxygen and carbon dioxide. the complex three-dimensional (3-D) anatomical structure of the airways and vasculature accompanied by the regional and functional variations in airway and vascular caliber on the basis of position, oxygen, and ventilation demand and a host of other physiological factors makes bioengineered reproduction of the lung simply challenging. In this regard, some key research areas in lung bioengineering can be identified with the aim of de novo generation of tissues and organs for transplantation where the native organ's structure and function are recapitulated (74, 127, 130, 144, 171, 183, 205). The major themes include 1) the need for scaffolds that mimic lung architecture, allowing for 2) repopularization of such scaffolds with appropriate lung cells that will importantly recapitulate the many aspects of lung function while 3) permitting sufficient vascularization for both maintaining viability of the engineered organ and oxygenation/ventilation, and concurrently 4) minimizing immunogenicity and rejection of the organ.

The challenge of maintaining architecture.

One of the major aspects of lung architecture that make reproduction so difficult is the need for a comparably complex scaffold upon or within which the functional elements of the lung can reside and, importantly, respond to changing demands by maintaining mechanical properties critical for the transplanted lung to function (74, 142, 144, 171, 183, 184, 205). There are now several experimental approaches to generating scaffolds to integrate the airway and vascular structures necessary to make a functional lung, including engineered scaffolds that use synthetic materials and microfluidic networks, and novel fabrication and printing technologies, and as decellularized scaffolds from human or large-animal lungs that use the native extracellular matrix structure for support (Fig. 1). Whether the structural and functional properties of these scaffolds, particularly once recellularized, sufficiently recapitulate those of the native lung is not clear, and represents an area of increasing focus. Some questions within this landscape then become: 1) what components make a good scaffold for recellularization? 2) how do cell-matrix (or more broadly cell-scaffold) interactions influence the organization, structure, and function of cells within the bioengineered lung? and 3) what are the mechanical properties of such scaffolds, and how do they influence the function of the new lung? The current status of research relating to these questions is addressed in the section Materials, Matrix, and Mechanobiology.

Fig. 1.

Bioengineering the lung. Current thoughts regarding how best to engineer a lung for transplantation in humans involve a 2-pronged approach. The first is to generate an anatomically appropriate lung scaffold, followed by recellularization of the scaffold. One approach would be to chemically and enzymatically decellularize lungs from large mammals, including humans, thus creating a scaffold containing many of the native extracellular matrix (ECM) proteins such as laminin, collagens, and fibronectin in an anatomically appropriate format. The other approach is to use the basic building blocks of the ECM or other biocompatible materials [such as Matrigel and polylactic-coglycolic acid (PLGA)] to generate biocompatible scaffolds that can then be engineered into an anatomical model. Decellularized or acellular scaffolds could then be reseeded with a number of cell types of varying pluripotency such as embryonic stem cells (ESCs), progenitor cells (e.g., those derived from fetal cells), mesenchymal and induced pluripotent stem cells (MSCs and iPSCs, respectively), or even adult lung cells. The idea would be for such cells to eventually entirely occupy the surfaces of the decellularized scaffolds, facilitated by a bioreactor environment with ventilation (mechanical stimuli), perfusion, and additional growth factors and stimuli to generate a functional transplant-ready lung. ASM, airway smooth muscle; PASM, pulmonary artery smooth muscle.

The challenge of recellularization.

As with other bioengineered organs, introduction and management of appropriate cell types to repopulate lung scaffolds represent a separate and evolving challenge as we understand the importance of cell selection; the potential and breadth of their regenerative capacity; their lineage and vintage; the methods for their introduction; the patterns of adherence, growth, migration, and interaction with the local environment including elements of the scaffold itself; their immunogenic potential; their fate; and, importantly, their function. Substantial research has explored the use of autologous primary cells or those with various level of pluripotency [e.g., embryonic stem cells (ESCs), bone marrow-derived and mesenchymal stromal cells (MSCs), and lung-derived progenitor cells, induced pluripotent stem cells (iPSCs)] to seed airway and vascular components in an engineered lung (Fig. 1). In this regard, the current understanding of factors that control stem or progenitor cell adhesion, migration, differentiation, and functionality is still relatively immature. Additionally, an unanswered question is whether mature cells of the lung (or other organs) can be induced to develop pluripotency, or at least regenerative capacity for recellularization. The current status of research, and the potential limitations of the approaches being considered, are explored in the section Recellularization in Lung Bioengineering.

The challenge of mechanobiology.

Rhythmic lung inflation and deflation that accompanies breathing imposes mechanical forces at micro and macro scales that not only tax the cells and matrices of the lung, but also determine and modulate cell growth, survival, and site- and context-appropriate function per se. Similarly, normal blood flow, pulse pressure, and shear forces induced by flowing blood are critical for vascular homeostasis. Added to these passive forces are the active mechanical forces imposed on both airways and vasculature by luminal contraction and relaxation in response to changing demands, and by endogenous and exogenous factors such as local and circulating agonists and antagonists. Although in silico models (79, 83, 99, 132, 148, 196), in vitro and ex vivo systems (59, 92, 103, 108, 118, 176, 193, 202, 203), and whole-organ approaches (94, 143, 205, 206) have been developed to explore the importance of such mechanical forces, further advances in airway, lung parenchyma, and vascular mechanobiology will be critical for bioengineering a lung capable of withstanding the internal and external forces exerted on the transplanted lung within the chest cavity and will be critical to ensuring that lung function occurs without airway collapse, injury, or failure of gas exchange. The challenges to our current understanding of lung mechanobiology and their implications for a bioengineered lung are discussed in Materials, Matrix, and Mechanobiology.

The challenge of vascularization.

Even with appropriate scaffolds and cells to repopulate, one of the major limitations in bioengineering of most organs has been the lack of a vascular system that matches the metabolic and functional needs of the organ being recapitulated. Here, it is critical that a vascular system capable of oxygenating tissue is available and fully functional immediately after transplantation, particularly for the lung. A consistently patent, low-pressure vascular system that mimics the native pulmonary vasculature is needed to respond to blood flow and metabolic demands, and to avoid pulmonary hypertension, right ventricular failure, thromboses, and other complications that are known to occur with allogeneic transplantation. Again, as with the airways, structural integrity in the face of internal and external mechanical forces is important. Thus understanding interactions among mechanical properties of bioengineered vascular structures, hemodynamics, coagulation, and gas exchange is as important as on the airway side of the bioengineered lung. Admittedly, this has been one of the major limitations in lung bioengineering (partly reflecting the complex architecture as well as relationships between airway and vascular function), a topic discussed in Putting It All Together.

The challenge of immunogenicity.

An important consideration (as with allogeneic transplants) is the immunological response of the transplant recipient to the bioengineered lung (102, 131, 189, 190). Regardless of the materials used to engineer a lung, generalized immunosuppression will likely be the initial approach to minimizing rejection, although techniques to enhance acceptance of a bioengineered lung containing specific types of scaffold materials and cell types are obviously a major research focus [e.g., biocompatible materials, novel (ideally personalized) sources for stem cells, and cell differentiation techniques].

Certainly, a multimodal approach involving biocompatible materials and scaffolds, cellularization of implantable structures (especially using stem cells and other cell types with potential for redifferentiation and proliferation), and/or guided regeneration of native tissues forms the basis for the considerable ongoing research and attempts at clinical applications. Such tissue engineering requires synthetic or natural matrices to support and guide cell attachment and growth to eventually recapitulate the native organ's structure and tissue-specific biological functions. However, at the present stage, a bioengineered lung that is ready for transplantation is still a research ideal that faces substantial challenges on many fronts. Nonetheless, the considerable recent and ongoing research in lung bioengineering provides an outstanding platform for understanding basic lung physiology and pathophysiology, for developing and characterizing novel models of lung disease, and for enhancing overall tissue engineering toward transplantation.

Materials, Matrix, and Mechanobiology

A major approach to lung bioengineering has been the use of scaffolds generated by decellularization of whole organs. Decellularization involves removal of native cells from the extracellular matrix (ECM) of tissues to produce a 3-D organ scaffold (7, 10, 28, 37, 57, 70, 125, 171, 184, 205) that ideally retains the 3-D anatomy of the native organ. Whole-organ decellularization using perfusion techniques was initially successfully demonstrated for a whole-heart scaffold (142), and since then for multiple organs including lung, liver, kidney, and pancreas using organs derived from rodents through humans (11, 12, 20, 60, 119, 139–141, 153, 156, 167, 177, 181, 194, 207). Decellularization involves the use of detergents, salts, enzymes, and physical approaches to removing cells while preserving not only ECM composition and organ architecture, but also importantly, ECM bioactivity and mechanics, with <50 ng of dsDNA/mg tissue remaining. A number of decellularization methods have been reported (7, 10, 56, 57, 171, 205), although an optimal decellularization technique is likely organ and context dependent. Given the intricate architecture of the human lung and the importance of the alveolar-capillary interface, this is certainly an advantageous but challenging approach toward developing an elastic scaffold of sufficient surface area for gas transfer.

Decellularization and ECM.

A key aspect of lung decellularization is that native ECM composition must be retained while cellular debris is sufficiently removed. This is a major challenge. For example, several studies have demonstrated that ECM proteins such as collagen, laminin, elastin, and fibronectin can be retained following decellularization, whereas others have reported reduced levels of ECM proteins and important growth factors (1, 27, 152, 162, 217). Here, the protocol for decellularization appears to be important. For example, decellularization of rat lungs using the zwitterionic detergent CHAPS (153) preserves airway and vascular structures with retention of collagen and, to a lesser extent, elastin, whereas glycosaminoglycans (GAGs) are depleted. On the other hand, such matrices can be successfully reseeded with lung epithelial and microvascular endothelial cells and do show gas exchange. In comparison, decellularization with SDS also maintains alveolar and vascular structures and permits recellularization and oxygen exchange (141) and even survival of a rodent transplanted with such a bioengineered lung. However, elastin content is also reduced. On the other hand, decellularization of mouse lungs with Triton X-100 and sodium deoxycholate with DNase (156) allows for retention of collagen, elastin, and GAGs. Furthermore, with this approach, laminin is retained (albeit at a lower level than in the native lung), which would be important for differentiation of alveolar cells following recellularization, and thus eventual barrier function. Conversely, the Triton X-100-based approach found lower levels of GAGs, which may be helpful in reducing adverse immune stimulation, but this remains to be proven. However, the relative advantages of this approach vs. those used by other studies in the overall context of recellularization and eventual function upon transplantation remains to be determined. Such comparisons likely are further driven by the relative roles of different ECM elements in cell attachment, differentiation, migration, immunogenicity, and other aspects of cell-matrix interactions that are in fact not completely understood.

Mechanics and the decellularized lung.

Altered ECM composition and structure affects the strength and mechanics of the remaining scaffold. Decellularization influences matrix stiffness due to both removal of cells and the altered ECM composition, but the effects are not consistent. For example, depending on the method of decellularization, the same organ can show increased tangential modulus and stiffness (142, 218), or minimal to no change (215). Consistently, decellularized lung scaffolds have been evaluated by use of pressure-volume curves (141, 153, 156) and show hysteresis with decreased compliance (increased stiffness), attributed to removal of surfactant and altered ECM (39). However, as mentioned above, the technique of decellularization appears to play a role in the resultant ECM composition.

Modulatory factors in the decellularized scaffold.

Although an important aspect of the ECM is that it provides structural integrity and mechanical stability for the various forces within the lung, it is now clear that the ECM also harbors a number of proteins, enzymes, and growth factors that are important for signaling of epithelia, smooth muscle, nerves, and blood vessels (38, 100, 114, 155). Furthermore, the composition of the ECM changes with age, from early postnatal development through aging. In turn, these ECM components are themselves modulated by diseases such as asthma, COPD, and pulmonary fibrosis (i.e., many of the same diseases that are indications for lung transplantation). Factors such as altered oxygen levels (2, 16, 69, 110, 120, 157, 210, 221, 226), environmental influences (23, 211), and inflammatory and profibrotic mediators (3, 4, 9, 30, 154, 209, 229) can modulate the expression of collagens, fibronectin, elastin. and other important lung ECM components. Indeed, there is now increasing evidence that mechanical forces per se can, on the one hand, influence ECM composition (200), whereas on the other, the ECM influences the cellular properties in the lung (33, 88, 134, 164, 173, 195). Mechanical properties and forces can further modulate cell fate, particularly the differentiation of stem cells on the basis of mechanotransduction (48, 49, 149, 161, 204). Accordingly, the source of the native lung from which a scaffold is derived likely alters the initial conditions under which a lung is subsequently bioengineered. The importance of these initial conditions on the final “product” is still under investigation, given the many permutations and combinations of scaffold materials and cell types being used. Indeed, the relationships between ECM composition and lung cellular properties are themselves being investigated in the context of basic lung biology and alterations in pathophysiological conditions.

At the level of the bronchial epithelium, mechanical forces can be manifest at the luminal level via airflow and ciliary transduction with the additional influence of mucus. However, in addition, mechanical forces can occur through bronchoconstriction as well as the cyclical stretching of the airways that occurs with rhythmic breathing (66). Here, data suggest that mechanical forces can result in activation of the EGF receptor pathway (89) that can then promote airway remodeling: an aspect likely influenced by the ECM (113, 151). Conversely, cyclical stretching of the epithelium can modulate prostanoid synthesis (169). Furthermore, dynamic compressive strain of the epithelial layer that is associated with bronchoconstriction of a circular airway can promote epithelial properties such as solute transport, viral gene delivery (199), and secretion of ECM-modifying proteins such as tissue factor (147). The mechanisms involved in mechanotransduction are still under investigation, but they could include Rho kinase (197), among other factors, which is in turn influenced by the ECM (154).

There is substantial interest in the cell-matrix interactions of airway smooth muscle (ASM) in the context of diseases such as asthma (231) that are likely extensible to the growth of airway cells in a bioengineered or decellularized scaffold. For example, mechanical stretch of human ASM induces TGF-β1 expression through mechanosensitive pathways such as RhoA/ROCK and even PI3/Akt (121). Such effects may be important in the pattern of ASM growth on a decellularized scaffold. Furthermore, molecular pathways such as RhoA are important in maintenance of contractility (96) and may thus play a role beyond recellularization. Here, the matrix stiffness may be relevant. For example, collagen-based hydrogels softer than average airway tissues promote secretion of vascular endothelial growth factor (VEGF) by human ASM cells, whereas stiffer gels stimulate cell proliferation and decrease contractility (i.e., they promote a proliferative cell phenotype) (176). These relationships between extracellular elastic modulus and cellular function are correlated with changes in integrins and other signaling proteins. From a functional standpoint, cyclical forces on the smooth muscle may modulate contractility (65).

Linkages between matrix stiffness and fibroblast properties also appear to be important (103, 108). In human fibroblasts, yes-associated protein (YAP) and transcriptional coactivator with PDZ-binding motif (TAZ), transcriptional effectors of the Hippo pathway, have been recently reported as key matrix stiffness-regulated coordinators of fibroblast activation and matrix synthesis (103). Here, YAP and TAZ are prominently and preferentially expressed in fibroblasts that populate fibrotic (presumably stiffer) human lungs, but not healthy lung tissue. Both YAP and TAZ are activated when fibroblasts are grown on pathologically stiff matrices but not physiologically compliant matrices. Importantly, suppression of these effectors attenuates matrix synthesis, contraction, and proliferation on pathologically stiff matrices. Together, these results identify YAP and TAZ as mechanoresponsive factors that amplify and sustain fibrosis. Similarly, Rho/ROCK and Mkl1 have been implicated in fibroblast mechanoactivation (77, 179, 233). These emerging albeit limited data show the importance of cell-matrix interactions within the airway wall, which may be important in the context of repopulating a bioengineered scaffold.

Given the importance of mechanical forces of ventilation on the alveoli, understanding mechanotransduction and mechanical forces at the alveolar level is obviously of importance (32, 80, 150, 186, 192, 193), especially if mechanical forces alter ECM generation by alveolar cells. For example, even brief periods of alveolar stretch can induce an inflammatory cascade (227) that can in turn influence alveoli and the pulmonary vasculature. Here, the extent of stretch is likely important (8, 40, 163, 195), and stretch effects may involve a range of pathways including Fas/FasL (93), Rac1 (41), PLA2 (116), GEF-H1 (17), caveolins (222), and Rho kinase (42), with modulating effects of oxygen (78, 164), ROS (31), and inflammation per se (72, 82).

At the vascular level (33, 135, 228), cessation and alterations of blood flow can be mechanosensed by the pulmonary endothelium—an aspect that is highly relevant to conditions such as pulmonary embolism, bypass surgery, and extracorporeal membrane oxygenation, and organ harvest and revascularization in the context of lung transplantation. Cessation of blood flow can be sensed by endothelial caveolae via PECAM-VEGF receptor-VE cadherin, which further leads to cell depolarization, increased intracellular Ca²⁺, activation of NADPH oxidase, and generation of reactive oxygen and nitrinergic species (33, 224). These cascades alter vascular tone and neovascularization on the one hand, but on the other, they also lead to oxidative injury and/or inflammation and cell death under pathophysiological conditions (33, 224). Although the scaffold will ultimately permit repopulation with a variety of lung cell types to facilitate gas exchange, an important aspect will be endothelial and epithelial barrier function to prevent edema. Here, the role of the ECM per se is not well understood. However, in addition to well-known endothelial barrier signaling pathways, there is recent evidence that the glycocalyx plays a role in barrier functions that may be modulated by mechanical forces (43, 87, 115, 136).

Mechanotransduction can also occur at the level of the vascular smooth muscle; for example, following arterial stiffening in the context of hypertension. Here, the resultant high pulsatility flow can induce endothelial dysfunction (involving decreased eNOS and increased vasoactive mediators and cytokines) that leads to changes in the smooth muscle. In the presence of excitation-contractions, high pulsatility flow results in smooth muscle hypertrophy (and contractile proteins) but not proliferation, whereas mechanical forces on the smooth muscle per se reduce their contractile potential (but do not alter proliferation) (172). Understanding these key events is important not only in the context of allogeneic lung transplantation, but also in the design of vascular scaffolds and networks in the bioengineered lung to ensure proper vascular structure and function.

Although the many elegant studies summarized above have yielded important information on how decellularization techniques influence the resultant scaffold and the potential modulatory factors that influence the resulting scaffold that would be used for eventual recellularization, some fundamental, unanswered questions remain: 1) what is the most effective balance between removal vs. retention of ECM elements in the scaffold in the context of recellularization? 2) perhaps more importantly, how should this balance be determined for eventual recapitulation of lung mechanics and functionality in the recellularized lung? 3) with the assumption that each donor lung (regardless of species) has a somewhat different ECM portfolio that is further modulated by environmental and disease exposure, what is the best approach to individualizing decellularization? and 4) conversely, should decellularization be customized on the basis of the target recipient or the sources from which the cells that will be derived during the recellularization process? Obviously, the answers to these questions lie much further away than the current focus on decellularization, but they will be important to consider for the decellularized lung approach to be eventually useful for a transplantable lung.

Biocompatible Lung Scaffolds

As discussed above, one ideal approach for a bioengineered lung would involve the use of nonimmunogenic, biocompatible materials of appropriate mechanical properties that facilitate recellularization and overall functionality that are key to the transplanted organ. An alternative approach would be degradable biomaterials that facilitate de novo generation, secretion, and deposition of “native” scaffolds by the cells initially introduced into the biomaterial, thus minimizing long-term immunogenicity. Both of these approaches are being explored in the context of lung bioengineering. Several biocompatible and biodegradable scaffolding materials have been explored for the engineering of 3-D lung tissues. Initial and obvious choices are natural ECM proteins such as collagen (29, 71, 145, 146, 191). However, due to its low mechanical properties in loose hydrogel configurations, collagen by itself is not load-bearing unless it is cross-linked or mechanically compressed, which may separately be relevant to the issue of reduced recellularization potential, and reduced flexibility of the scaffold for eventual use. On the other hand, addition of soluble elastin to collagen hydrogels improves the mechanical properties of the resultant scaffold, with further enhancement by seeding with lung fibroblasts that can introduce a variety of ECM proteins that result in stiffness equal to that of a single alveolar wall (≈5 kPa) (44). Such innovative approaches involving a mix of native lung ECM components may provide initial building blocks toward regeneration of new functional lung tissues.

In addition to collagen, Matrigel (101, 123, 124) and other natural polymers (6) have been explored in both in vitro and in vivo studies. Synthetic polymers such as polyglycolic acid (175), polylactic-coglycolic acid (PLGA) (122), and pluronic F-127 are also considered of substantial interest in the context of lung scaffolds (35). An obvious advantage of using such synthetic material is the ability to tailor their chemical and biological properties with additional manipulation to achieve desirable mechanical properties. For example, pluronic-based scaffolds seeded with lung progenitor cells, inject able polymers, and collagen-GAG scaffolds have been used to mimic alveolar structures (6, 34, 35) and allow in vivo implantation (35). However, realization of such approaches toward stable, implantable structures is still investigational. For example, a macroporous matrix from a blend of hydroxyethyl methacrylate-alginate-gelatin has been shown to support human lung epithelial cell proliferation for several weeks and to recruit cells while remaining biocompatible in vivo, but apparently it can also induce infiltration of immune cells (178). Nonetheless, such approaches hold significant potential for reseeding with cells such as alveolar epithelial cells and generating a functioning lung.

In terms of conducting airways, the creation of transplantable trachea has already been demonstrated clinically (106). However, the generation of structurally stable small bronchioles using biocompatible matrices is still under investigation. Here it is increasingly clear that cellular characteristics and responses are likely different in vitro in a 2-D environment as opposed to the 3-D topographies as would exist in vivo. Techniques such as electrospinning are attractive methods for producing 3-D topographies for culturing cells such as ASM because the spun fibers have dimensions that are comparable to those within the ECM of the airway. For example, an electrospun scaffold using nondegradable polyethylene terephthalate has been used to generate the topography of human ASM cells to demonstrate adhesion, alignment, and spindle-shaped morphology, thus generating fully aligned sheets of smooth muscle that demonstrate contractility to bronchoconstriction (126). On the other hand, the idea of initially matrix-free 3-D ASM constructs have also been investigated. For example, microfabricated tissue gauges have been used to develop a 3-D culture model of ASM that is capable of contractility, possesses strong cellular organization, and contains actin stress fibers, particularly with the inclusion of 3T3-fibroblasts (225).

Clearly, compared with the more substantial, albeit preclinical, efforts toward decellularization of mammalian lungs, research on de novo creation of biocompatible lung scaffolds is even more nascent. However, even as this landscape becomes better defined, it would be important to consider two fundamental questions: 1) what is the most appropriate portfolio of ECM and associated proteins required for efficient and effective recellularization? and 2) as with decellularization techniques, how should this portfolio be individualized or customized for recellularization based on cell sources, target recipient, and eventual recapitulation of lung mechanics and functionality?

Recellularization in Lung Bioengineering

Recellularization of a previously cellular, now-decellularized scaffold, or one newly generated using biomaterials involves introducing cells onto and into the matrix; promoting attachment, migration, proliferation; a customization of the ECM environment; and importantly, functionality that mimics the native tissue. A number of studies in different organ systems have now shown that native cells can attach, migrate, and proliferate in decellularized scaffolds (see Refs. 50, 171, and 214 for reviews). For example, transplanted decellularized aortic valves or large-vessel scaffolds can be endothelialized and recellularized by cellular migration upon reseeding. Clinical trials for trachea, bladder, and cardiac valves and vessels have been recently reported (67, 84, 100, 111). However, unlike these simpler organs containing a few cell types in predominance, the complex 3-D structure of the lung and its many functions makes use of decellularized lungs challenging, whereas the large number of cell types with various functions on the other hand also make it a challenge to recellularize in a spatiotemporally and functionally appropriate manner. Nonetheless, decellularized lung scaffolds have been used to understand differentiation and patterning of stem cells and progenitor cells, albeit in small animals. In terms of cell types, the focus to date has been on epithelial cells and endothelialization.

Alveolar epithelial cells.

Given the important functional requirement for a reseeded lung to oxygenate and ventilate, there has been appropriate focus on the alveolar epithelium, which consists of two major types: squamous alveolar epithelial type I (ATI) cells, which occupy a majority of the alveolar surface area in the native lung, and surfactant-producing cuboidal type II (ATII) cells, which occupy a small area but constitute >65% of the epithelial cell population (105). Transformation of ATII cells to ATI cells upon reseeding onto an acellular human alveolar matrix was demonstrated a number of years ago (105). Fetal cells (typically derived from the native organ of interest, and in nonhuman species) have been reported in several studies for tissue bioengineering and seeding of scaffolds (156, 174, 182). An advantage is that fetal cells can retain phenotypic markers as well as their functionality, and, relevant to epithelial and other typically polarized cells, their spatial or compartmental orientation. For example, decellularized rat lung scaffolds seeded with fetal or neonatal rat lung cells are capable of gas exchange (141, 153). Intratracheal seeding of fetal alveolar epithelial cells into decellularized lung (156, 174, 182) has been explored, with the scaffold supporting growth of such fetal ATII cells (156) as evidenced by markers such as pro-SP-C or cytokeratin 18. Thus it appears that the decellularized scaffold retains components necessary to direct the differentiation of cells with proliferative and differentiating capacity. Interestingly, such properties of the scaffold may be tissue-specific in that alveolar progenitor cells seeded onto a liver scaffold do not show lung-specific proteins such as SP-C or aquaporin-5, which are alveolar-type (AT) markers.

It is certainly much easier to acquire adult lung cells (e.g., from surgical resections, transplant recipients, or donor lungs deemed unfit for transplantation), and such adult cells have been explored for their potential for reseeding of scaffolds (alveolar epithelial cells, fibroblasts, and small airway epithelial cells) (131, 137, 194). However, as expected, these adult cells show restricted proliferative capacities and are therefore unlikely to be useful for reseeding large surface areas needed for a bioengineered human lung.

Endothelial cells.

As discussed above, endothelial integrity and function will be key for gas exchange and regulation of lung fluid status, as well as ensuring that sufficient cellular proliferation is maintained to replenish the endothelial barriers. In rat lung scaffolds, use of rat lung microvascular endothelial cell or human umbilical vein endothelial cells (HUVECs) for reseeding the vascular compartment has been shown to result in inadequate and uneven re-endothelialization (141, 153, 182). This can obviously promote thrombosis and is likely insufficient for barrier function.

Embryonic and mesenchymal stem cells.

Stem cells and progenitor cells are of obvious interest on the basis of their ability to expand in culture and differentiate into multiple cell lineages, including ATII, bronchial epithelial, or pulmonary endothelial cells (18, 36, 52, 97, 160). However, the use of stem cells requires extensive knowledge about the spatiotemporal mechanisms underlying lung development, which is relatively well understood in mouse models and to a lesser extent in humans (90, 223), raising the question about the approach to be taken in the use of human-derived stem cells to induce lung cell lineages. Nonetheless, given the cellular and spatial complexity of the lung, it is unlikely that a single lung stem cell is capable of generating all the necessary lineages for a bioengineered lung. Indeed, even in the nonembyronic lung, it is more likely that stem cell or progenitor cell lineages of different regenerative capacities are present in spatially and temporally restricted patterns. A good example is that lung epithelial cells do have the capacity to proliferate and replace lost cells (e.g., during injury caused by environment, mechanical ventilation, etc.) (22, 47). In this regard, adult stem cells are thought to represent a potential cell source for lung repair (14, 24, 81, 106, 109, 130, 133, 168). Furthermore, 3-D scaffolds reseeded with lung progenitors and stem cells, with additional stimuli via growth factors and mechanical forces (simulating those in the native lung) have been shown to induce cellular differentiation and formation of anatomically appropriate lung tissue (122, 166). Human embryonic stem cells (ESCs) seeded onto macaque lung scaffold slices express lung markers (128). Mouse ESC differentiation has been compared between lung scaffolds and hydrogel matrices such as gelfoam, Matrigel, and collagen I (36), and expression of lung markers and scaffold-facilitated cellular differentiation into epithelial and endothelial lineages have been found. However, a problem with ESCs is that the decellularized scaffold may have limited ability to provide all of the necessary cues for differentiation into the range of cell types required. This may be an even greater issue in decellularized scaffolds derived from adult lungs previously exposed to disease or environmental factors that substantially alter the native ECM. On the other hand, predifferentiation of such ESCs can enhance differentiation once seeded onto lung matrix (129). For example, predifferentiation of mouse ESCs into ATII-like cells facilitates their seeding onto decellularized lungs and neovascularization in situ (85). The relationships between ESC differentiation and scaffold properties are still under investigation, and will likely be context-specific in that ESCs or their derivatives may require specific differentiation cues to promote site-specific cell types.

Mesenchymal stem cells.

Mesenchymal stromal, or stem cells (MSCs), are quite appealing in the context of organ engineering on the basis of our current ability to isolate them from peripheral tissues such as bone marrow or the fat of patients, to expand their population, and then redifferentiate them into other cell types in culture. This raises the potential for readministration of autologous MSCs for patient-specific organ engineering (112). However, realization of this potential will be dependent on the ability of MSCs to recapitulate both anatomical and functional features of the different lung cell types. There is now considerable evidence that MSCs play a role in regeneration and repair (61, 62, 81, 91, 112, 201) and in secretion of growth factors, immune modulators, and other factors that support the idea that MSCs may be better taken up in organ scaffolds (158, 198, 220), especially if scaffolds enhance their differentiation (86, 219), presumably through cell-matrix interactions (64, 161, 174, 232). Several studies have shown that MSCs can attach and proliferate in lung scaffolds (39, 117, 131, 137). For example, MSCs derived from human adipose tissue and bone marrow on decellularized rat lungs express markers of pulmonary epithelial cells (117), although the source of the MSCs may play a role in their local behavior such as cell attachment, migration, and proliferation. MSCs within lungs have also been identified that have the ability to express markers of ATI and ATII cells (73, 95, 214). In this regard, a particularly relevant finding is that MSCs may be less finicky toward the scaffold milieu in terms of attachment (19, 21, 170, 180, 213, 217). However, what is not yet established is whether MSCs fully redifferentiate into specific lung cell types within a scaffold and recapitulate the functions of the native cell type.

In addition to ESCs and MSCs, iPSCs have been considered in lung recellularization (55) and repair (14). These cells are generated by reprogramming of somatic cells toward an embryonic state (230) and therefore are highly amenable for the use of patient-derived cells. Recent studies have used iPSCs for lung recellularization and shown that human iPSCs can differentiate to ATII cells that attach and proliferate on decellularized lung matrices while maintaining their cellular phenotype (55, 58).

Immune cells.

An interesting problem that has been scarcely addressed is how to recapitulate the many immune functions of the native lung. Beyond serving as a barrier to the many environmental and biological insults the lung normally faces, lung cells such as the epithelium initiate and modulate immune responses, whereas conversely, the lung hosts a range of immune cells, including dendritic cells, macrophages, and others. Although it is expected that host immune cells would take up residence in an engineered lung and initiate immunity, balancing their effects on maintenance of immune function vs. initiating rejection and dysfunction of the transplanted lung will likely be an important issue to address in the future.

Overall, regardless of the approach taken to providing a scaffold, there is clearly substantial interest and exciting progress in the arena of recellularization techniques utilizing a range of cell types. Whereas the optimal approach to recellularization has yet to be defined, it is indeed possible to ensure attention to some important questions toward realizing a functional recellularized lung: 1) is there a single cell source that will eventually be optimal for repopulating a scaffold with the variety of different cell types needed for functional recapitulation? 2) does the cell source need to be matched to the type of scaffold used (a particularly relevant question if decellularized scaffolds can modulate cellular phenotype and function)? 3) for a scaffold, what are the best routes for recellularization toward functional airways vs. vasculature? and 4) what is best way to balance immunogenicity and host immunity vs. recapitulating the immune function of the native lung?

Putting Things Together

An interesting aspect of recellularization of the lung is that on the one hand a scaffold can be seeded through the airway route, or via the vasculature, but on the other hand limiting cells destined for forming blood vessels to the vasculature rather than to the airway can be a challenge. Several studies have shown that seeding lung cells such as epithelial cells and fibroblasts via the tracheal route and HUVECs via the pulmonary artery allows for reconstitution of alveolar-capillary membranes that can perform gas exchange (25, 141, 153, 156, 182). The ultimate goal is to achieve complete recellularization by distributing cells throughout the branches of the airway and vasculature, which is currently still under investigation. Nonetheless, it is also clear that for clinical success an engineered lung will be dependent on establishing a stable, functional blood supply, especially given that denuded vasculature in scaffolds is highly thrombogenic (11, 140, 165). These issues raise the question of how to best prepare the engineered lung for eventual transplant. Since the first reports of functioning recellularized tissue engineered lungs (141, 153) transplanted into rats showing sustained gas exchange, there has been much excitement regarding the potential of clinically applicable bioengineered lungs. However, the current state of the art in whole-lung bioengineering is admittedly at a proof-of-concept stage, with substantial questions and roadblocks remaining. As highlighted above, issues relating to the type and properties of the lung scaffold and cells for reseeding are particularly important. However, even once these issues are addressed, the question remains as to how to maintain and optimize the engineered lung prior to transplantation. Here, the area of bioreactors has undergone tremendous scientific and technical development, particularly given the incentives for studying lung regenerative strategies in a physiological environment, and for preserving or reconditioning explanted lungs for transplantation. In this regard, bioreactor technology ranges from small-scale apparatuses to study lung cells in vitro and gas exchange, through microfluidic vascular networks mimicking the alveolar-capillary membrane (74–76, 188) to whole-lung bioreactors that facilitate decellularization of the lung by perfusion (thus maintaining vascular structures) followed by recellularization in sterile, closed systems with subsequent ventilation and perfusion (143, 144). However, there are still several technical and logistical hurdles in terms of determining the appropriate parameters for ventilation vs. perfusion based on the scaffolds and cells involved, provision of nutrients, and the eventual functionality of the reseeded lung (143). Here, many of the topics discussed above, such as cell type, scaffold ECM, and mechanical properties, play important roles. Additional issues when it comes to human transplantation will include the need to generate sufficient numbers of appropriate cell types to ensure complete recellularization and inhibited immunogenicity—issues that potentially may be addressable with the use of iPSCs. On the issue of immunogenicity, a particularly challenging roadblock may be the balance between minimizing rejection and resultant failure of the transplanted, bioengineered organ vs. recapitulating and maintaining the inherent immune functions of the lung and its resident cells (or additionally in the context of infectious and environmental exposures).

Conclusion

Although lung bioengineering is admittedly in a nascent stage, there has been significant progress in understanding the important aspects of biocompatible matrices that will allow for reseeding of cells to generate appropriate cell lineages that will adhere, proliferate, migrate, and function in the appropriate locations within the complex anatomy of the lung. Here, use of iPSCs and other cell types may allow for diminished to absent immunogenicity with sufficient cell quantities. Thus, in the foreseeable future, recellularization-based bioengineered lungs may prove to be an effective and alternative clinical therapy for treatment of recalcitrant lung diseases that currently require allogeneic lung transplantation.

GRANTS

Support for this study was provided by National Heart, Lung, and Blood Institute Grants R01 HL-056470 and HL-088029 to Y. S. Prakash, R01 HL-092961 to D. J. Tschumperlin, and R01 HL-114887 and P01 HL-014985 to K. R. Stenmark.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y.S.P., D.J.T., and K.R.S. conception and design of research; Y.S.P., D.J.T., and K.R.S. analyzed data; Y.S.P., D.J.T., and K.R.S. interpreted results of experiments; Y.S.P. prepared figures; Y.S.P. drafted manuscript; Y.S.P., D.J.T., and K.R.S. edited and revised manuscript; D.J.T. and K.R.S. approved final version of manuscript.