Engineering Tissues from Induced Pluripotent Stem Cells

Stem cells hold tremendous promise for replacing or regenerating tissues damaged by injury and disease as well as to study developmental biology and pathomechanisms. Early studies suggested the potential of various multipotent cell populations derived from adult humans, such as mesenchymal stromal cells (MSCs; also referred to as mesenchymal stem cells and bone marrow stromal cells ¹ ), to form tissues derived from all three embryonic germ layers. ² However, subsequent study proved that although MSCs and other adult-derived stem cell populations can have some regenerative benefits, ³ they are multipotent rather than pluripotent. Importantly, rigorous analysis showed that they cannot differentiate into some key lineages, including heart muscle cells (cardiomyocytes) that are required to repair tissues with limited regenerative capacity.^4,5

The discovery of methods to culture human embryonic stem cells (hESCs ⁶ ) offered the promise of generating organ- and tissue-specific cell types in a controlled laboratory setting for applications in tissue repair and in vitro disease modeling. Our extensive and growing knowledge of development biology has provided guidelines for specific cues to use to guide cells down specific lineages. ⁵ However, studies with hESC are ethically controversial and strongly regulated because these cells are derived from human fetal tissue. Yamanaka's discovery of methods to reprogram somatic cells into induced pluripotent stem cells (iPSCs)^7,8 circumvented the ethical conundrums of hESC research. Human induced pluripotent stem cells (hiPSCs) offer the potential for personalized tissues, which can be used for regenerative medicine and/or in vitro studies to tailor other medical interventions. ⁹

Providing cells with an appropriate physiological three-dimensional (3D) environment is important for basic disease modeling and for engineering grafts to replace damaged tissues. Tissue engineering approaches that involve growing cells on 3D scaffolds have proven invaluable to maximizing survival and therapeutic potential of transplanted cells. In many cases, cells gain emergent properties when they are cultured in the appropriate 3D niche environment, enhancing the power of in vitro models to predict how specific drugs or genetic conditions can affect patient health. In the past decades, a variety of approaches to generate 3D tissues have been introduced and employed for different applications: they were typically based on various types of scaffolds ranging from decellularized matrices and electrospun fiber mats to natural and synthetic hydrogels.^10–13 Modern approaches now focus on the generation of microphysiological systems (MPSs) by harnessing developmental biology and self-assembly processes to generate organoids or by combining microfabrication techniques with tissue engineering to generate organ-on-a-chip systems.^14–18 MPS enables the recapitulation of even more complex in vivo processes and interactions in an in vitro setting. This special issue of Tissue Engineering Part A, entitled “Engineered tissues derived from induced pluripotent stem cells for disease models, drug discovery, and replacement therapies,” brings together a diverse set of study from nine leading research groups in the field. These original research articles focus on the potential of tissue engineering approaches to enhance our ability to use hiPSC to generate physiological in vitro models, to model inherited disease and injury in a controlled in vitro setting, and to empower our ability to regenerate damaged tissues.

The ability to model inherited diseases such as cardiomyopathy is a key advantage of genetically defined iPSC. Because mechanical cues are critical to cardiomyocyte (CM) biology, tissue engineering approaches that control the mechanical environment of iPSC-derived CMs are critical for our understanding of disease pathophysiology. ¹⁹ An approach that combines decellularized extracellular matrix (ECM) from adult heart with disease-prone hiPSC-CMs is described by the Tung group (in this issue). The authors used established methods to decellularize adult left ventricular tissue from adult pigs to provide an in vivo-like microenvironment for hiPSC-CM derived from a patient with the inherited arrhythmogenic cardiomyopathy (AC). ²⁰ AC is a disease involving mutations of the intercalated disk, a specialized structure that reinforces cell–cell adhesions in heart and skin. Importantly, the engineered heart slice model used by the Tung laboratory caused AC-prone hiPSC-CM to upregulate several markers of CM maturation, including TNNI3, PLN, and KCNJ2, compared with when these same cells were cultured in monolayers. Conduction velocity is a key predictor of the potential of CM populations to generate arrhythmias, and, importantly, the AC-prone hiPSC-CM cultured on engineered heart slices exhibited a significantly lower conduction velocity than hiPSC-CM derived from healthy patients, described in a previous report by the same authors. ²¹ Arrhythmias could be induced in a robust manner with point pacing in hiPSC-CM on engineered heart slices, opening the way to use this model to study antiarrhythmic drug therapies. As in studies by other groups, culturing hiPSC-CM in an engineered tissue format increases the baseline conduction velocity of the conglomerate of cells, ²² making it easier to distinguish healthy from disease-prone cell populations and potentially facilitating discovery of basic disease mechanisms.

Differences in physiology and molecular biology between humans and animals have been cited as a major cause of failure during drug development, ²³ necessitating the use of human in vitro models. However, iPSC-derived tissue cells, including hiPSC-CM, are immature, exhibiting a phenotype more reminiscent of cells in a fetus than a fully developed adult. Tissue engineering and bioreactor approaches have been used to coax these cells into exhibiting more mature structure and function.^{14,15,22,24,25} Chen and Vunjak-Novakovic (in this issue) have used a combination of tissue engineering with prolonged (2 week) culture to mature hiPSC-CM-based tissues. In this study, they demonstrate that this initial maturation improves the robustness of an in vitro model of ischemia-reperfusion injury (IRI). IRI describes the clinical observation that although reperfusion of occluded arteries is the most effective clinical approach to treating infarcted myocardium, the reperfusion itself leads to more cell death and ultimately affects the size of infarcts. In their study, these authors first demonstrated that only matured tissues exhibited a full range of molecular responses to IRI (including lactate dehydrogenase release) that have been identified in animal studies. As in previous study, this suggests that even relatively simple tissue engineering approaches to improve maturation in hiPSC-derived tissue cells substantially enhances the power of those cells for modeling disease and injury. Interestingly, preischemic conditioning, which reduces the extent of IRI in animal models but is not a viable clinical option, ²⁶ decreased the extent of IRI in the engineered tissue model, whereas an antioxidant small molecule, N-acetyl-l-cysteine, failed to ameliorate IRI.

A key facet of tissue engineering for improving maturity of hiPSC-CM is that these approaches involve coculturing CMs with stromal cells. However, the actual source of these cells has been different from one laboratory to the next, varying from embryoid-body-derived stromal cells to fetal cardiac fibroblasts.^14,27 McDevitt and colleagues have conducted an in-depth systematic analysis of the molecular phenotype of several well-defined stromal cell populations on the single-cell level. Part of the importance of using single-cell approaches to studying stromal populations is that the specific molecular markers used to define “fibroblasts” are not well defined compared with markers used to define other cell populations (e.g., iPSC or T cells). This makes it difficult to do precise molecular characterization on bulk populations obtained by sorting marker-positive cells from a heterogeneous pool. These authors identify differences in cell shape, cytoskeletal morphology, and gene expression among different stromal cells. They further link stromal cell identity to formation of heterogeneous tissues made with a combination of stroma and purified hiPSC-CM. Importantly, formation, along with calcium-handling and gene expression of heterotypic tissues made by combining stromal populations with purified iPSC-CMs. Interestingly, although these cells do not play a developmental role in forming the heart, dermal fibroblasts appeared to be one of the most optimal stromal populations for enhancing the calcium handling of hiPSC-CM.

A common issue of engineered tissues is that with increasing complexity of the recapitulated biological systems and processes, the model itself and the conduction of experiments becomes increasingly complicated and user dependent. This frequently entails transferability problems limiting application and commercialization. In this issue, Schneider et al. present a user-friendly microfluidic organ-on-a-chip platform, which integrates hiPSC-derived cardiac tissues with microscale dimensions in a vasculature-like perfusion. Their heart-on-a-chip platform allows for the injection of cells and generation of 3D tissues by means of centrifugal forces utilizing solely standard laboratory centrifuges and easily adoptable preparation routines. The authors demonstrate the capability of the system to generate multiple scaffold-free cardiac tissues in a parallelized manner using a minimal number of cells due to very high loading efficiencies. The platform was then utilized for a functional characterization and proof-of-concept compound screening highlighting the generation of independently beating replicates with a physiologically relevant structure and function. In addition, the authors provide an improved open-source software for quantitative analysis of bright-field video microscopy of cardiac tissues.

Despite extensive efforts to make hiPSC-derived tissue cells mature, there may be inherent limitations on our ability to force cells grown for short timeframes in vitro to undergo programming that takes decades to progress in the body. ²⁸ However, cellular engineering may allow us in some cases to “shortcut” the maturation process by artificially modulating the levels of genes that increase or decline with aging. In this issue, Acun and Zorlutuna use CRISPR/Cas genome editing to create hiPSC deficient in the α-subunit of hypoxia inducible factor 1 (HIF-1), a transcription factor that declines in expression in endothelial populations of aged animals. ²⁹ Within a robust 3D-synthetic ECM (gelatin-methacrylate)-based lumen formation assay, HIF-1α-deficient hiPSC-derived endothelial cells (hiPSC-ECs) showed impairment in the ability to form new lumens under hypoxic, but not normoxic conditions. Compared with their isogenic counterparts, HIF-1α-deficient hiPSC-ECs were sensitized to mitochondrial reactive oxygen species accumulation and loss of viability under hypoxic conditions.

The promise of hiPSC for cell-replacement therapies remains enticing. However, two obstacles that have long been associated with cell-replacement therapy are (1) obtaining the right cell population and (2) controlling the biology of this cell population after transplantation, remain. Phenotypic characterization, in addition to analysis of molecular markers (e.g., fluorescence-activated cell sorting characterization of surface antigens) is a critical component to understanding cells' quality and therapeutic potential. In this issue, Zoldan and colleagues demonstrate the ability of hiPSC-derived endothelial progenitors (hiPSC-EPs) to create vascular networks in collagen gels that mimic the in vivo environment into which these cells will eventually be transplanted for therapeutic revascularization. Importantly, the authors have developed an unbiased, semiautomated approach to quantifying vascular network formation by hiPSC-EP, and used this to elucidate how material properties of the ECM microenvironment influence vascular network formation. Previous studies have shown key differences in the network forming potential of pluripotent stem cell derived and primary human EC populations, ³⁰ and new analytical tools such as the network analysis approach described here by Zoldan and colleagues will be useful in optimizing in vitro treatments of hiPSC-EP and to enhance their therapeutic potential.

Given the need to enhance cellular maturation and improve cell function and tissue organization, bioreactor strategies have previously been incorporated to enhance engineered tissues.^24,31 Dynamic application of forces have been used in vitro to induce physiological ECM alignment to direct cell organization and improve function. ³² In this issue, Smith and colleagues use a stretch bioreactor to induce alignment and elongation of hiPSC-derived cortical neuron (hiPSC-CN)-like cells in vitro. Stretching monolayer hiPSC-CN grown on extensible transparent films and subjected to different uniaxial strain rates, they were able to grow robust axon tracts up to 1 cm in length in <1 month. The stretch approach, thereby, provided a much quicker response and growth than more typically employed anisotropic confinement approaches based on microfabricated patterns. The authors envision their hiPSC-based engineered functional cortical axons could provide an autologous approach for cerebral axon transplantation after injury, as a potential treatment to prevent persistent neurological and cognitive deficits.

An inherent challenge in organ transplantation and cell replacement therapy is controlling transplanted cell function. Cellular and tissue engineering approaches have been used to control transplanted cell biology, as demonstrated in studies by Corrales and colleagues and Guilak and colleagues (both in this issue). The earliest studies involving hiPSC-derived tissue cell implantation in humans have involved retinal pigment epithelial cells (RPEs). HPSC-derived RPE transplanted as monolayer sheets showed excellent performance and safety biocompatibility in nonhuman primate studies and in initial clinical application.^33,34 In the preclinical case study described here, Corrales and colleagues have used a nanostructured fibrin-agarose biomaterial to provide the correct alignment and organization of hiPSC-RPEs upon transplantation into mouse and pig models. Interestingly, hiPSC-RPE viability was dramatically reduced when these cells were precultured on scaffolds during cellular maturation, whereas cells remained viable when they were matured before being combined with the scaffolds. Transplanted human hiPSC-RPE in murine and porcine hosts remained localized at the site of implantation and sensitive assays for human-specific genetic elements suggested that these transplanted cells and their genetic material remained localized at the implant site in the eye.

The host inflammatory response presents a significant obstacle for regenerative medicine approaches to repair many tissue types, including cartilage. In addition to biomaterial-based tissue engineering, synthetic biology approaches that involve engineering transplanted cells to overcome host inflammatory responses are a second promising and localized approach to solving the problem of transplanted cell viability. Guilak and colleagues describe a lentivirus-based synthetic gene circuit system wherein mouse iPSC-derived engineered cartilage constructs are engineered to secrete an anti-inflammatory molecule, interleukin-1 (IL-1) receptor antagonist (IL-1Ra, anakinra), in response to inflammatory signals that activate an nuclear factor-kappa B (NF-κB) response element. Because the transplanted cells both receive the inflammatory signals that activate NF-κB and secrete the anti-inflammatory IL-1Ra, the system is localized and self-regulating. The authors show the ability to use existing biomaterials [e.g., poly(ɛ-caprolactone)] to facilitate viral gene delivery to cells. Although lentiviral delivery makes the absolute dosage of factor less controlled than it would be with nuclease (e.g., TALEN or CRISPR/Cas)-based germline genome editing, this lentiviral approach and the coupled biomaterial-based delivery could be readily translated to introduce rationally designed gene circuits to induce cell-based anti-inflammatory or other beneficial signaling through a variety of pathways and effectors.

Altogether, these articles provide impressive examples of burgeoning research that harnesses cellular and tissue engineering principles to enhance the power of stem cell technology. In the future, hiPSC-based engineered tissues will be key aspect for the understanding and treatment of human diseases.