Engineering Biomaterials with Micro/Nanotechnologies for Cell Reprogramming.

Although Spemann initially proposed the concept which would later be coined as somatic cell nuclear transfer (SCNT) in 1938, it was not until 1952 that Briggs and King successfully developed the technique and demonstrated that introducing the nuclei of a blastula cell into an enucleated egg could give rise to a cleaved blastula. 8 Ten years later, Gurdon transferred the nuclei from differentiated Xenopus intestinal epithelial cells into enucleated eggs, and this resulted in the formation of fully functional tadpoles. 39 These findings revealed that the reversal of cell differentiation, now known as reprogramming, was feasible, which initiated mammalian cloning experiment using SCNT in 1996 and the derivation of SCNT-derived human embryonic stem cell lines in 2013. 40 , 41 In comparison to SCNT and cell fusion, cell reprogramming offers a more robust approach through the ectopic expression of defined transcription factors ( Figure 1 ). The concept of reprogramming regained wide interest in 2006 when Takahashi and Yamanaka demonstrated that they were able to generate iPSCs from mouse somatic cells via the combined expression of four transcriptions factors, Oct4, Sox2, Klf-4, and c-Myc (OSKM). 9 It has been demonstrated that these iPSCs closely resemble embryonic stem cells (ESCs) in terms of karyotype, stemness phenotype, and differentiation capacity. This significant breakthrough highlighted the possibility of generating any desired cell type from a somatic cell through a combination of iPSC reprogramming and directed differentiation, and numerous pioneering studies have unraveled the mechanisms for cell reprogramming. 42 - 44 Although human ESC-derived cells have been primarily used in clinical trials, iPSCs are considered superior to ESCs without the ethical concerns involving the destruction of embryos. It is encouraging that some preliminary clinical trials have recently launched using human iPSC cells (hiPSCs)-derived cells to treat diverse diseases, including macular degeneration, aplastic anemia, heart disease, endometriosis, Parkinson’s disease, and spinal cord injury. 45 - 49 However, iPSCs can be associated with genetic and epigenetic abnormalities that may arise during the reprogramming process. 50 As a result, there is an increased risk of tumorigenicity and immunogenicity when using iPSC-derived cells. Moreover, some additional hurdles, including unstable differentiation, low conversion efficiency, and lengthy differentiation still exist, thereby limiting the clinical translation of iPSC-based cell therapies. To circumvent these issues, scientists have explored other methods of manipulating cell fate ( Figure 1 ). Direct reprogramming is an alternative strategy that directly converts fully differentiated somatic cells into distinct cell types using transcription factors, microRNAs, or chemical compounds. 10 In comparison to iPSC reprogramming, this method is considered faster and safer as cells do not proceed through an intermediate pluripotent state. The direct reprogramming was reported in the late 1980s that the forced expression of MyoD, a transcription factor in the skeletal muscle lineage, could induce mouse embryonic fibroblasts (MEFs) into stable myoblasts. 51 Since then, numerous studies in the past two decades have shown that this method can be applied to obtain a wide variety of target cell types, including mouse or human hepatocytes, 52 , 53 neurons, 54 - 56 pancreatic β cells, 57 , 58 renal tubular epithelial cells, 59 immune T cells, 60 embryonic Sertoli-like cells, 61 endothelial cells, 62 and cardiomyocytes. 63 - 66 Direct reprogramming is a promising approach for personalized medicine as it has been shown that patient-derived cells recapitulate the disease phenotype, which offers great potential for disease modeling, drug screening, and personalized treatments. For example, direct neuron cell reprogramming has been used to study human disease progression in vitro, such as Alzheimer’s disease. 67 Although direct reprogramming can be utilized to generate cells with a mature phenotype, this may not be ideal for clinical translation. Thus, more recently, direct reprogramming has been applied to generate various types of expandable stem cells or progenitors, such as neural stem cells, 68 oligodendrocyte precursor cells, 69 hepatic stem cells, 70 hematopoietic stem cells, 71 , 72 nephron progenitor-like cells, 73 intestinal progenitor cells, 74 skeletal muscle progenitor cells, 75 , 76 and cardiac progenitor cells. 77 , 78 By providing extensive proliferative capacities and restricted differentiation potentials, these directly reprogrammed progenitors have been shown to be more engraftable and desirable for transplantation in regenerative therapies. 71 In addition, it is worth noting that directly reprogrammed cell populations usually have cellular heterogeneity. Thus, cell purification is commonly required to isolate the target cells from the rest of the population using fluorescence-activated cell sorting (FACS), 71 immunopanning, 69 or other sorting techniques. Indirect reprogramming has been demonstrated as another reprogramming approach. This process entails converting somatic cells through a transient induction of a partially reprogrammed cell using the Yamanaka factors, followed by differentiation into the target cell type. 79 Similar to direct reprogramming, this indirect approach does not give rise to pluripotent cells and thus bypasses the teratoma-forming potential that is typically associated with iPSCs. Using this indirect reprogramming method, several types of lineage-specific cells and multipotent progenitors have been generated from mouse and human cells. 80 - 83 To date, most reprogramming studies have been conducted in vitro. Alternatively, several studies have shown that resident cells can be directly reprogrammed into another desired cell type in situ, such as mouse pancreatic cells, 84 - 86 neurons, 87 - 93 oligodendrocytes, 94 cardiomyocytes, 95 - 97 hepatocytes, 98 , 99 retinal cells, 100 epithelial cells, 101 and lymphocytes, 102 through the overexpression of lineage-specific transcription factors, microRNAs, small molecules, and other reprogramming factors. Such studies have shed light on an emerging field in exploiting in vivo lineage reprogramming to mend the broken tissues by harnessing and converting internal cells into targeted functional cells without cell transplantation, risks of contamination, and tumorigenesis. 14 Notably, the in vivo microenvironment offers an optimal niche for enhancing direct cell reprogramming. Similarly, iPSC reprogramming has also been investigated in vivo . 103 - 107 It has been reported that iPSCs derived in vivo demonstrate totipotency and a closer resemblance to ESCs in comparison to iPSCs generated in vitro . 105 In contrast to the finding that DNA damaged and aged cells display lower reprogramming efficiencies in vitro, tissue injury and senescence promotes in vivo OSKM-mediated reprogramming. 106 , 107 Although significant advances have been made for in vitro cell reprogramming during the past decade, in vivo reprogramming is still in the nascent stages. Moreover, numerous obstacles related to reprogramming efficiency, quality, safety, and controllability must be overcome before there is robust translation of reprogramming therapies. Promisingly, engineering proper delivery systems and microenvironmental cues via micro/nanotechnologies is expeditiously promoting vital breakthroughs in the field of cell reprogramming.

Cell reprogramming is a highly valuable method for engineering cell fate. This comprehensive review highlights the recent progress in engineering biomaterials with advanced micro/nanofabrication techniques to provide delivery platforms and biophysical and biochemical cues for enhancing in vitro and in vivo cell reprogramming. However, many challenges remain to be addressed before the clinical application of reprogramming technologies, including efficiency, quality, safety issues, in situ targeting, and the underlying signaling mechanisms. To address these challenges, multidisciplinary collaborative research from the fields of materials science, bioengineering, biology, and medicine would be required. Optimized delivery systems need to be developed or selected from a diverse group of drug- and gene-delivery strategies to achieve the desired reprogramming outcome. First, reproducibility and higher efficiency will be the critical criteria in developing the optimal delivery methods for cell reprogramming. For in vivo reprogramming, if specific parts of the body, organs or tissues, or a subpopulation of cells are targeted for cell type conversion, a drug-delivery system incorporated with a specific ligand, antibodies, peptides, or aptamers can be developed. Second, as small molecules offer powerful tools to manipulate cell fate by inducing or enhancing cell reprogramming, the co-delivery of nucleic acids with small molecules or the multidelivery of chemical cocktails can be developed to promote cell reprogramming in vitro and in vivo . Third, smart and stimuli-responsive delivery systems can be custom-designed to achieve more efficient in situ targeted delivery of reprogramming factors, which is an emerging area of biomaterials research. Fourth, biophysical and biochemical cues can also be incorporated into the drug-delivery system to enhance cell reprogramming, yet the optimal conditions for reprogramming remain to be clarified. Biophysical manipulation provides a promising approach to regulate cell fate and improve the reprogramming process. First, the majority of biophysical cues (including topography, stiffness, and extracellular force) and biochemical cues (including ECM, soluble factors, and small molecules) have been employed for cell reprogramming individually, which demonstrates the feasibility of these approaches and helps elucidate the underlying mechanisms. Combining such cues may synergistically enhance reprogramming efficiency, but it is a challenging task to figure out the optimal conditions with so many possible combinations, crosstalk, and potential side effects. In addition, temporal regulation of the reprogramming process at the early or late phase may require different microenvironmental factors. If sufficient experimental data can be collected and specific experiments can be designed to address the combinatory effects, a machine learning approach can be used to optimize the reprogramming conditions and protocol. Second, to make it relevant to in vivo conditions, reprogramming in 3D scaffolds with desired and integrated cues will be a key step to generate and apply reprogramming-mediated engineered tissues in vitro or in situ. Third, whether and how biophysical factors may influence in vivo reprogramming is not well understood and deserves further investigation in the future. Advances in other cutting-edge technologies will continue to expand the frontiers of cell reprogramming. First, microfluidic platforms enable the delivery of reprogramming factors in a more controlled and efficient way from the perspectives of composition, dosage, and spatiotemporal features, which can also offer well-defined nanotopographical and mechanical cues, including fluid shear stress, cyclic strain, and compression. It is expected that the orchestrated interplay of the controlled delivery and combined physical and chemical cues can further enhance cell reprogramming. Meanwhile, the microfluidic environment allows directed generation of functional cells in the same platform, which creates a model of “tissue/organ reprogramming-on-a-chip”. Second, 3D printing holds great promise for engineering tissues and organs for disease modeling and regenerative medicine. 269 3D printing enables the manipulation of the microenvironment with spatially defined mechanical, structural, and chemical properties, thus, making it suitable to generate 3D complex niches for cell reprogramming. Third, single-cell sequencing analysis will provide insight into the heterogeneous cellular responses and fate determination during the reprogramming process 270 , 271 and help elucidate the signal pathways, molecular identity, and underlying mechanisms by which biophysical and biochemical cues regulate cell reprogramming. Fourth, CRISPR/Cas9 genome-editing platforms provide powerful tools not only for genetic correcting of hereditary diseases 272 - 274 but also for somatic cell reprogramming, as demonstrated in recent studies for the derivation of iPSCs 275 and neurons. 276 Viral-free CRISPR/Cas9 delivery systems have shown promising results, which may enable more in vivo applications. Hence, simultaneous cell reprogramming and CRISPR/Cas9 gene editing, through the co-delivery of both gene-editing systems and reprogramming factors, may guide the design of next-generation cell-reprogramming technologies. The advancement and integration of cell reprogramming and biotechnologies present exciting opportunities for research and technology translation in these interdisciplinary frontiers. We expect that the rational design and development of delivery systems and biomaterials that control microenvironmental cues will accelerate the translation of reprogramming technologies for cell engineering and therapies.

Various genetic and epigenetic regulators or reprogramming factors have been utilized for cell reprogramming ( Figure 2A ), including transcription factors in the form of proteins, 11 , 108 episomal plasmids, 109 , 110 messenger RNA (mRNA), 111 micro-RNA (miRNA), 112 , 113 small molecules, 67 , 114 , 115 mini-circle DNA, 116 , 117 synthetic transcription factors, 118 and CRISPR-Cas9 gene-editing tools. 119 It is important to note that although these reprogramming factors can yield differences in the reprogramming efficiency and reproducibility, this can potentially be improved upon using an optimal delivery system. 120 , 121 Thus, the choice of a safe and effective delivery system is paramount in achieving successful reprogramming with specific reprogramming factors. 122 , 123 Current delivery modalities for cell reprogramming can be classified into three groups: biological, physical, and chemical delivery systems ( Figure 2B ), and their comparisons are summarized in Table 1 . Herein, we will focus on the recent advances of physical and chemical delivery systems, considered as biomaterials-based approaches in a broad sense, for cell reprogramming. Viral delivery of reprogramming factors relies on genetically engineered viruses to transduce transcription-factor-encoding plasmids into cells, including integrative vectors based on retrovirus and lentivirus and nonintegrative vectors based on adenovirus, adenovirus-associated virus (AAV), and Sendai virus. 64 , 124 Recombinant viral vectors are often preferred for cell reprogramming because of their high transduction efficiency. However, viral gene delivery usually has limited DNA packaging capacity, unpredictable gene dysfunction, cytotoxicity, immunogenicity, and tumorigenicity. Detailed information on virus-based delivery for iPSC and direct cell reprogramming can be found in several excellent review papers. 11 , 14 , 120 , 125 , 126 Physical delivery strategies involve directly delivering reprogramming factors into the cytoplasm or the nucleus via membrane-disruption methods, including electroporation and mechanical disruption ( e.g., cell squeezing and nanostructure penetration). Electroporation has become one of the most popular nonviral gene-delivery techniques since initially reported in 1982. 127 Bulk electroporation (BEP) is commonly conducted with millimeter- to centimeter-sized chambers, where the cell suspension and molecules to be delivered are mixed in a conducting buffer solution between two electrodes. Electrical pulses are applied to induce transient permeability of cell membrane and diffusion/endocytosis for cytosolic cargo delivery. 128 Electroporation-based methods can deliver a diverse range of cargos, including nucleic acids or genes (small interfering RNAs, miRNAs, mRNA, DNA plasmid), proteins or protein complexes, small molecules, and nanoparticles. This multifaceted capability provides the potential for a multitude of in vitro and less amenable to in vivo applications, such as gene editing, 129 adoptive immunotherapy, 130 - 132 regeneration medicine, 133 - 135 cell reprogramming, 136 , 137 and others. 128 , 137 , 138 BEP-based delivery has been used to deliver episomal vectors to generate integration-free hiPSCs in vitro 109 , 139 and to induce direct reprogramming into neurons 140 and oligodendrocytes in vivo . 94 Nevertheless, despite its simplicity, BEP-mediated transfection has drawbacks, including excessive cell damage caused by high electric fields, highly stochastic transfection, and low efficiency of intracellular delivery (diffusion dominated). 141 , 142 To overcome these challenges, microfluidic-based electroporation technologies, such as microchannel electroporation (MEP), have been developed to operate at lower voltages, enabling the enhancement in efficiency and viability, 143 single-cell 143 or high-throughput molecular delivery, 144 and precise dose control. 138 Recently, another vortex-assisted microfluidic electroporation system, named microscale symmetrical electroporator array ( μ SEA), has been developed with integrated micropatterned planar electrodes in the vortex cell-trapping chamber and equipped with real-time visualization functionality ( Figure 3 ). Such a system not only allows for sequential intracellular delivery of various molecules but also presents several other benefits, including dosage control, real-time visualization, multimolecular delivery, high transfection efficiency, and improved cell viability. 123 However, the cargo size of the MEP approach is still limited because its cytosolic cargo delivery relies on diffusion and endocytosis. With advances in nanofabrication and engineering techniques, recently, nanoelectroporation (NEP) systems have been developed for efficient cell transfection. 145 For example, a nanochannel electroporation device was designed with a microchannel–nanochannel–microchannel structure, and single cells were precisely positioned at the tip of the nanochannel via optical tweezers ( Figure 4A ). In contrast to other electroporation methods, NEP can transfect cells faster and more efficiently at the single-cell level with a precise dosage of transfection agents ( e.g. , oligodeoxynucleotide, siRNA, quantum dots nanoparticles, GFP plasmid), while not affecting cell viability. 141 However, it is hard to trap adherent cells by optical tweezers, and it is even less likely suitable for trapping large numbers of fast, adherent cells ( e.g. , MEFs) for cell reprogramming. Therefore, a more advanced high-throughput NEP biochip has been further developed by the same lab whereby cells were loaded using a centrifugal spinning method ( Figure 4B ). This technique greatly increased the cell loading efficiency from 10 to 80% compared to optical tweezers and allowed for a large number ( i.e. , 10 000 cells for a device) of MEF cells to be transfected with OSKM plasmid for mouse iPSC reprogramming. 146 After NEP transfection, more than 90% of cells were alive and the transfection efficiency was significantly higher than that with BEP. However, further systematic characterization of specific gene and protein expressions of transfected cells should be performed to determine the final iPSC reprogramming efficiency and elucidate additional advantages of using nanoelectroporation for in vitro cell reprogramming. In addition, the NEP techniques also enable tissue nanotransfection (TNT) of nonviral cargos for rapid, highly effective, and deterministic in vivo reprogramming ( Figure 5 ). The TNT platform could efficiently promote the reprogramming of mouse skin cells into neurons with the delivery of Ascl1/Brn2/Myt1l (ABM) plasmids. This approach elicited 50- to 250-fold greater gene expression in vivo compared with conventional bulk electroporation. In particular, via the delivery of Etv2/Foxc2/Fli1 (EFF) plasmids, skin cells could be reprogrammed into endothelial cells to increase vascularization and rescue skin tissue and whole limbs under necrotizing ischemia. Moreover, the TNT platform could also mediate RNA-based transfection of microRNA302/367 cluster to induce pluripotency of skin. 136 Another nanotransfection approach is nanostraw electroporation, which can achieve high-throughput cell transfection with dosage control and high uniformity. 147 Although the nanoscale electroporation platforms have displayed more efficient and enhanced transfection capabilities, some challenges remain to be addressed, including precisely targeted cell delivery in vivo and complex fabrication protocols that require special expertise for the manufacturing of miniaturized electroporation systems. Mechanical disruption is an alternative method to disrupt the cell membrane for instantaneous delivery. 122 , 148 Unlike electroporation methods, mechanically triggered cell disruption does not depend on electric fields, external materials, endocytosis, or cargo properties. 122 This approach is particularly applicable to difficult to transfect cell types. Recently, the advances of microfluidics and nanotechnology have enabled mechanical-disruption approaches, such as cell squeezing and nanostructure-based penetration for cell reprogramming. Cell squeezing permits the rapid deformation of cells flowing through a narrow microfluidic channel that is half to one-third of the cell’s diameter in size, thus leading to mechanical disruption of the plasma membrane that allows for the diffusion of various materials of interest into the cytosol. 149 - 152 Particularly, microfluidic-mediated cell squeezing has been used to deliver OSKM recombinant proteins into human fibroblasts for iPSC reprogramming and showed a 10-fold improvement in colony formation compared to electroporation or cell-penetrating peptides ( Figure 6A ). 152 To disrupt both the plasma membrane and nuclear envelope, electric-field-driven transport was combined with cell squeezing in a microfluidic device for the enhanced delivery of plasmid DNA ( Figure 6B ). 144 A major advantage of the cell squeezing method is the use of simple devices to transfect diverse cargos into large numbers of cells within a short time period. However, current cell-squeezing methods have a critical limitation regarding inconsistent transfection efficiency for asynchronous cell populations with different sizes. Thus, more advanced squeezing devices need to be designed with a certain range of geometric constrictions to disrupt cells of different sizes. 122 In addition, several nanostructure-based penetration methods have demonstrated the potential for effective high-throughput delivery of an assortment of biomolecules into living cells, including nanowires, 153 nanoneedles, 154 nanospears, 155 and nanostraws. 147 , 156 The nanoscale devices can penetrate the plasma membrane, thus enabling the straightforward delivery of different biomolecules, loaded onto the surface or within the bulk structure, into the cytosol and nucleus. 137 It has been demonstrated that porous silicon nanoneedles can efficiently deliver nucleic acids into the cytosol (greater than 90% efficiency), and in particular, in vivo neovascularization can also be achieved using the nanoneedle-mediated minimally invasive and localized delivery of angiogenic VEGF-165 gene ( Figure 6C ). 157 Nanostraw is another alternative cell penetration approach to mediate cytosolic delivery of different biomolecules. Unlike other nanostructure-mediated delivery methods, microfluidic systems can be integrated with nanostraw platforms to allow permanent fluidic access and spatial and temporal control in delivery. 158 Additionally, electroporation systems can be further combined to generate nanostraw electroporation devices ( Figure 6D ). 147 These devices have been shown to greatly enhance plasmid transfection (~81%) for Chinese hamster ovary cells, who display relatively low (5–10%) transfection rates using nanostraw-mediated plasmid delivery. As such, this nanostraw electroporation system offers a powerful, high-throughput transfection platform with excellent dose, spatial, and temporal control, as well as high efficiency and cell viability. Although nanostructure-mediated delivery techniques remain in the early stage of rapid development, they have provided some exciting delivery platforms for a wide range of delivery applications and show great promise for the delivery of different reprogramming factors for in vitro and in vivo cell reprogramming. However, all the nanostructure-mediated penetrations usually lack a cell-specific targeting mechanism and cannot easily navigate 3D tissues, thus nanodevices are optimally suited for transfection into easily accessible 2D tissues such as the epidermis and retina. 101 Chemical delivery approaches, often involving nanoparticles or nanocarriers, are attractive in the drug- and gene-delivery field due to their suitable size for cell penetration, large surface area for loading molecular cargo, and protection of loaded substances. 159 As a carrier for nonviral gene delivery, nanoparticles are usually regarded a advantageous to viral transduction in terms of larger transgene delivery, low immunogenicity, ease of transfection, low genetic integration, and minimal insertional mutagenesis risks, 160 although the efficiency might be lower than that with virus-based methods. Nowadays, chemical-based nanocarriers, including lipids, polymers, and hybrid systems, have been widely applied to deliver plasmid DNA, miRNA, mRNA, recombinant proteins, and small molecules for diverse disease treatments 161 - 164 and have been utilized for the delivery of reprogramming factors for cell reprogramming. Lipids are amphiphilic molecules, consisting of hydrophobic tails and hydrophilic heads. As one of the most widely used nonviral gene carriers, lipid-based vectors/nanoparticles can deliver diverse cargos, including drugs, DNA, RNA, and genome-editing proteins into cells in an endocytosis-dependent manner. 165 - 168 Cationic lipids ( e.g. , DOTAP and DOSPA) contain three structural components: a cationic headgroup, one or two hydrophobic tails, and a linking group between these domains, which are the main components that bind with negatively charged nucleic acids or molecules to form the lipid-based lipoplexes. In addition, neutral lipids ( e.g., DOPE and DOPC) within the formulations can function as “helper lipids” to enhance nanoparticle stability and transfection activity. 161 Currently, several lipid-based vectors have been developed and applied to achieve cell reprogramming ( Figure 7A ). 111 , 113 , 169 , 170 Lipid-mediated repeated transfection was used to deliver expression plasmids containing the Yamanaka factors to reprogram MEFs into iPSCs, without evidence of plasmid integration, although the conversion efficiency was low ( i.e. , less than 0.0029%). 171 iPSCs also can be reprogrammed from human fibroblasts with enhanced efficiency through the delivery of a modified RNA cocktail using cationic lipid carriers. 111 In addition, lipid-based vectors can also be used to achieve in vivo cell reprogramming. It has been reported that the delivery of a single miRNA (miR-21) using DOPC/M-DOPE/Tween-80/DiO lipid nanoparticles can convert wound site macrophages to fibroblast-like cells in diabetic mice. 172 Similarly, mice cardiac fibroblasts can be directly reprogrammed into cardiomyocytes in vitro and in vivo via lipid-based delivery of modified miRNAs. 97 The miRNA-mediated reprogramming efficiency was significantly increased ( i.e. , by more than 10-fold) when combined with JAK inhibitor I. Therefore, further optimization of delivery methods that combine miRNA with small molecules for in vivo reprogramming may amplify their therapeutic implications. Polymer-based drug- and gene-delivery systems have been widely used to treat broad pathological conditions via less invasive or noninvasive routes. 166 , 173 The most commonly used polymeric carriers are nanoparticles because they offer advantages such as chemical flexibility, stable formulation, excellent biocompatibility, facile preparation, and controlled release capability. Recently, delivery strategies based on natural and synthetic polymers have been adopted for cell reprogramming ( Figure 7B ). Natural biopolymers like polysaccharides have been generally used as polymeric backbones for the formation of nanoparticles or as valuable drug- and gene-delivery carriers. 174 , 175 For example, a positively charged polysaccharide was employed to deliver plasmid mixtures (encoding Oct4, Sox2 , miR302/367) to generate hiPSCs from human umbilical cord mesenchymal stem cells. 176 Similarly, the direct reprogramming of mouse fibroblasts into neural cells can be achieved by co-delivery of a group of plasmids ( Ascl1, Brn4, Tcf3 ) with ethanediamine-modified Porphyra yezoensis polysaccharide (Ed-PYP) nanoparticles. 177 Besides natural polymers, cationic synthetic polymers have immense chemical diversity and unlimited potential for functionalization. 161 Several synthetic polymer nanoparticles have been used for cell reprogramming, such as those based on polyethylenimine, 178 poly( β -amino ester), 179 poly- (amidoamine) dendrimer, 180 and cationic bolaamphiphile. 181 Nontoxic cationic polyurethane–short-branch polyethylenimine (PU–PEI) nanoparticles have been developed to deliver Oct4/SirT1 plasmids into aged-related retinal pigment epithelium cells or light-injured rat retinas. 178 For example, the PU–PEI-mediated Oct4/SirT1 gene transfer converted the retinal pigment epithelium cells into progenitor-like cells, which rescued retinal cell loss and improved electroretino-graphic responses in light-injured rat retinas. On the other hand, biodegradable polyurethanes with functional groups can provide a more robust platform for subsequent diverse modifications and thus can be potentially utilized for drug and gene delivery for cell reprogramming. 182 Poly( β -amino ester) is another polymer that has been widely applied as a gene-delivery system. For example, it has been employed to deliver genes encoding disease-specific chimeric antigen receptors (CARs) in order to reprogram T cells for therapeutic purposes. 102 The poly( β -amino ester) carriers also exhibited higher transfection efficiency of delivering a single polycistronic plasmid (pCAG-OSKMG) for human iPSC reprogramming, when compared with commercial transfection reagents. 179 In addition, dendrimers such as poly(amidoamine) are promising nanocarriers due to their high loading capability, excellent biocompatibility, and functionalization capability. Positive charges on the terminal amine groups of poly- (amidoamine) dendrimers can promote gene loading, cell penetration, and endosomal escape during transfection, resulting in functional gene transfection. 180 Furthermore, mouse iPSCs can be generated from MEFs using a single plasmid (pOSKM) delivered through arginine-terminated polyamidoamine dendrimer (G4Arg) nanoparticles, which exhibited a transfection efficiency (~15%) higher than that of commercially available carriers, such as Lipofectamine 2000 (~5%) or FuGENE HD (~8%). 183 Taken together, synthetic materials not only enable the delivery of small molecules, proteins, and a variety of other compounds but also offer the potential for controlled nucleic acid transfection in vitro and in vivo , providing flexibility and possibility to manipulate cell fate. Next-generation drug- and gene-delivery systems with optimized kinetics will further improve targeted delivery and the efficiency of cell reprogramming. Inorganic and hybrid materials can serve as nonviral drug and gene carriers that generally possess versatile properties suitable for intracellular delivery, including diversity, multiple functionality, excellent biocompatibility, targeted delivery, and controlled release. 184 , 185 Recently, several reports have used inorganic and hybrid particles for cell reprogramming ( Figure 8 ), including mesoporous silica, 186 calcium phosphate, 187 , 188 graphene oxide, 189 , 191 and gold-based nanoparticles. 190 Mesoporous silica nanoparticles (MSNs) have been widely used as nanocarriers for gene and drug delivery, mainly due to their tailored mesoporous structure, high surface area ratio, biocompatibility, and functionalization capability. 192 One study reported that avidin-capped MSNs adsorbed with three transcription factors ( Ascl1/Brn2/Myt1l ) could directly convert mouse fibroblasts into functional dopaminergic neuronlike cells. Interestingly, the co-delivery of a neurogenesis inducer, ISX-9, in the MSNs further enhanced the reprogramming efficiency. 186 Therefore, this strategy of co-delivering small molecules and plasmids can also be applied to other delivery carriers to achieve enhanced reprogramming performance. In addition, calcium phosphate nanoparticles exhibited the efficient delivery of four plasmids encoding OSKM to generate iPSCs from human umbilical cord mesenchymal stem cells. 187 The delivered nucleic acids formed ionic complexes with cationic calcium ions through their electrostatic interactions, which were easily transported into cells via ion-channel-mediated endocytosis. Interestingly, the hiPSC reprogramming can be accelerated upon incorporating calcium-phosphate-based hybrid nanoparticles into 3D collagen scaffolds, benefiting from the controlled release of OSKM genes and the biophysical regulation of the 3D niche. 193 Additionally, gold nanoparticles have emerged as another promising platform for gene and drug delivery due to their high surface area, inertness, and easy fabrication. 194 Cationic gold nanoparticles (AuNPs) are efficient carriers because they can interact with nucleic acids as well as the negatively charged plasma membrane. 190 , 195 For example, it has been reported that in vitro and in vivo cardiac reprogramming can be achieved using hybrid AuNP-based nanoparticles. The AuNPs modified with arginine-rich peptide (RRRGYC) were mixed with GMT ( Gata4, Mef2c, and Tbx5 ) plasmids and polyethylenimine (PEI) to make the AuNP/GMT/PEI hybrid nanocomplexes, which promoted the cardiac reprogramming efficiency of human and mouse fibroblasts. More importantly, upon injection into infarcted mouse models, the nanocomplexes enhanced the in vivo cardiac reprogramming outcome and cardiac function. 190 Furthermore, hybrid materials such as graphene oxide (GO)–PEI complexes have been engineered to deliver mRNAs encoding OSKM for hiPSC generation. 191 Intriguingly, GO–Fe 3 O 4 –PEI complexes were designed to achieve stimuli-sensitive gene delivery for the reprogramming of peripheral blood mononuclear cells into hiPSCs. Notably, the cell transfection and reprogramming efficiency were significantly improved by the combined treatment of magnetic stirring and near-infrared photothermal stimulation. 189 Despite the fact that inorganic or hybrid nanoparticles are promising as reprogramming delivery carriers, there are still some critical issues that need to be addressed. First, the loading capability and stability of genes or drugs are limited to a certain extent by surface adsorption. The second shortcoming is the possible aggregation of the inorganic nanoparticles loaded with nucleic acids, which may reduce the transfection efficiency. Another problem is the excretion of inorganic nanoparticles, and their accumulation in cells may be harmful; thus the in vivo safety of inorganic particles should be rigorously demonstrated before clinical translation. Altogether, in addition to the reprogramming factors, the selection of an appropriate carrier among various biological, physical, or chemical delivery systems is critical to enable and enhance cell reprogramming. Undoubtedly, diverse drug- and gene-delivery strategies that have been developed for genome editing, tissue engineering, and disease therapies can also be tested and applied for cell reprogramming. 19 , 122 , 123 , 196 - 199 From the perspective of clinical application, using non-gene-integrating platforms to generate reprogrammed cells is preferable to those that integrate into the genome. Moreover, considering the advantages and disadvantages summarized in Table 1 and translation potential, chemical carrier-mediated delivery systems, especially multifunctional, stimuli-sensitive drug-delivery systems, 200 , 201 may be more promising for in vivo reprogramming therapies via the controlled local release of reprogramming factors and targeted delivery.

In addition to biochemical signals, biophysical factors are broadly involved in orchestrating morphogenesis, development, homeostasis, and tissue remodeling by modulating cell proliferation, migration, differentiation, lineage commitment, or apoptosis. 202 - 204 Cells can convert extracellular biophysical cues into intracellular biochemical signals via mechanosensing and mechanotransduction, which in turn can regulate chromatin organization, gene expression, protein translation, and cell function. 202 , 205 , 206 In the past decade, biomaterials and devices have been used to manipulate external biophysical and biochemical cues for cell reprogramming. 33 , 34 By using advanced polymerization chemistries and micro- or nano-manufacturing techniques, biomaterials with desirable physical and chemical properties can be designed and fabricated. 207 For elucidative purposes, we focus on the effects of microenvironmental factors, including micro/nanotopography, stiffness, extracellular force stimulation, 3D microenvironments, and biochemical modifications, on cell reprogramming ( Figure 9 ). Manipulation of micro/nano-scale surface topography has a profound influence on cell functions, such as cell morphology, adhesion, proliferation, migration, differentiation, and stemness maintenance. 208 - 210 Although extensive research has demonstrated that topography can direct cell differentiation toward specific cell lineages, 211 , 212 more recent studies have provided some insight into the role of topography in cell reprogramming. Biomaterial topography, in the form of parallel microgrooves and aligned nanofibers, significantly improved iPSC generation through alterations in cell shape that resulted in an increase in histone acetylation and methylation and mesenchymal-to-epithelial transition (MET) promotion. 213 Consistent with this observation, another study further confirmed that the MET process mediated topography-induced epigenetic changes during iPSC reprogramming from mouse somatic fibroblasts using graphene-coated substrates. 214 These substrates, which have two-dimensional honeycomb structures, yielded almost 4 times the number of positive mouse iPSC colonies compared to glass surfaces. Apart from iPSC reprogramming, micro-topography can also influence the direct reprogramming process. For example, microgratings promoted the direct reprogramming of fibroblast into induced neurons by modulating neurite branching, neurite outgrowth, and gene expression. 215 , 216 Similarly, the efficiency of direct cardiac reprogramming can be enhanced with topography modulation. 217 , 218 Parallel microgrooves were found to be capable of not only modifying the epigenetic state ( e.g., increased histone acetylation) but also regulating myocardin transcription factor A signaling to promote direct reprogramming into cardiomyocytes. Notably, microgrooved topography generated more mature cardiomyocyte-like cells, which may enhance functional performance after cell transplantation into an infarcted heart. 217 In contrast to microtopography, nanotopography has demonstrated a more influential role in inducing differentiation of hiPSCs to a neuronal lineage, 219 possibly due to the fact that these nanoscale features may more closely resemble the in vivo microenvironment. Upon examining how micro- and nanotopographical cues could influence direct reprogramming into induced dopaminergic neurons, it was determined that nanogrooved substrates further enhanced neuronal gene expression, maturation, and reprogramming efficiency through modulation of histone mark expression and MET. 220 However, whether nanotopography can regulate direct reprogramming into other cell lineages remains to be further explored. The underlying mechanisms of topographical influence on cell reprogramming are not fully understood. However, the mechanisms that have been elucidated for topography-mediated cell differentiation may also be applicable in reprogramming. For example, the aforementioned studies suggest that both micro- and nanotopographical cues can influence cell and nuclear shape to alter the epigenetic state, gene expression, and thus cell fate. 208 - 210 In addition, topography is strongly correlated to focal adhesion dynamics, leading to intracellular signaling transduction pathways triggered by topographic parameters of substrates. Moreover, topography-induced cytoskeletal changes can induce nuclear deformation that leads to altered nuclear mechanics and chromatin state, both of which can affect cell phenotype and function. 221 Substrate stiffness is one of the most widely applied physical cues to modulate stem cell differentiation. 222 - 224 For example, soft substrates promote the differentiation of human MSCs into neurons or adipocytes, whereas osteogenesis is favored on stiffer substrates. 225 Matrix stiffness can also modulate the cell-reprogramming process. Recently, it has been demonstrated that soft polyacrylamide hydrogels with a stiffness of 100 Pa enhanced mouse iPSC reprogramming from MEFs on the surface of the hydrogel, 226 whereas another study reported that the optimal stiffness for mouse and human iPSC generation was between 300 and 600 Pa in 3D polyethylene glycol (PEG) hydrogels. 227 Mechanistic studies have revealed that substrate stiffness can alter cell adhesion and cytoskeletal reorganization, via mechanotransductive mechanisms, to activate signaling pathways that modulate gene expression and, thus, cell behavior. 222 , 228 Although several intracellular mechanosensitive signaling cascades are regulated by matrix stiffness, such as AKT/YAP/RUNX2, 229 p190RhoGAP, 230 and YAP/TAZ mechanotransduction, 29 , 231 the underlying molecular mechanisms of how stiffness regulates cell reprogramming remain unknown and require further investigation. Extracellular mechanical forces can influence organ development and function in health and disease by triggering a series of cytoskeletal-mediated mechanotransduction and intracellular signaling cascades. 202 , 232 , 233 Recently, extracellular forces including shear stress, mechanical stretching, and electromagnetic forces have been used as effective biophysical cues to modulate cell reprogramming. Extracellular fluids, including blood plasma, interstitial fluid, and lymph, are crucial in the exchange of substances and the maintenance of the microenvironment. Fluid shear stress is an essential mechanical factor in flowing liquid that can induce cytoskeletal organization remodeling, trigger mechanosensitive gene expression, and alter cell behavior, demonstrating an important role in vascular homeostasis, vascular remodeling, cardiac development, atherogenesis, 234 , 235 and cancer metastasis. 145 , 236 In addition, fluid movement enhances convective mixing and transport. Compared to conventional adherent culture, faster and more efficient mouse iPSC reprogramming can be achieved in a stirred suspension bioreactor with retroviral transduction of transcription factors ( i.e., Oct4, Sox2 , and Klf4, with or without c-Myc ), displaying more than a 10-fold increase in efficiency compared to that with the adherent culture. 237 Fluid shear stress present within a dynamic culture suppressed differentiation and provided the additional benefit of selective aggregate formation, enabling reprogrammed cells to flourish. As a result, dynamic suspension-induced reprogramming is a promising and scalable approach for basic research and clinical uses. 238 Interestingly, 2D culture with dynamic mixing at the mid to late phase of reprogramming also demonstrates significant improvement of mouse iPSC reprogramming, which was found to be attributed to the enhancement of cell proliferation. 239 Mechanical stress and strain in vivo play fundamental roles in guiding cell behavior and tissue homeostasis. 240 At the cellular level, the mechanical stretch has demonstrated crucial regulatory effects on cell spreading, growth, lineage commitment, and stem cell differentiation on 2D substrates. 241 , 242 In addition, externally applied stretch (uniaxial, biaxial, and equiaxial) has been utilized to engineer functional and mature tissues in 3D scaffolds, such as in heart tissue, 243 , 244 skeletal muscles, 245 and blood vessels. 246 , 247 Mechanical stretch can also be used for in vivo tissue regeneration, for example, by activating hair stem cells and eliciting macrophage recruitment and M2 phenotype polarization to facilitate hair regeneration in a strain/duration-dependent manner. 248 Recent studies have shown that mechanical stretch has a positive effect on reprogramming. When dermal fibroblasts transduced with OSKM were seeded onto flexible membranes that were mechanically stretched, 3 and 5 times more human iPSC colonies were obtained upon equiaxial stretching for 4 days at 3 and 8% strain, respectively, than with the unstretched groups. 249 However, cyclic mechanical stretch did not improve the yield of direct cardiac reprogramming. 218 Therefore, the effects of mechanical stretch may be dependent on specific signaling pathways involved in the reprogramming process. In addition, various parameters of stretching such as mode, magnitude, rate, frequency, duration, continuity, and relaxing period may have profound effects on cell signaling and reprogramming, which is worth further studies. Moreover, the mechanical forces driven by electromagnetic or electric fields can improve cellular reprogramming. For example, extremely low-frequency electromagnetic fields (EL-EMF) significantly enhanced the reprogramming efficiency from mouse fibroblasts to iPSCs. EL-EMF induced dynamic regulation of epigenetic changes through the activation of a histone lysine methyltransferase. 250 In addition, electro-magnetized AuNPs provided an excellent interface to deliver EMF for cell reprogramming. Application of EMF to cells cultured on AuNPs directly reprogrammed fibroblasts into induced dopaminergic neurons both in vitro and in vivo. EMF stimulation specifically activated the histone acetyltransferase Brd2, leading to histone H3K27 acetylation and neuron-specific gene activation. 195 Furthermore, a triboelectrical stimulation system can be combined with nonviral gene delivery to directly convert fibroblasts into functional neuronal cells with higher efficiency. 251 The triboelectrical stimulation promoted dynamic changes of intracellular calcium levels in reprogrammed cells, including increased calcium influx and ERK1/2 pathway activation. These electromagnetic or electric stimulation systems can potentially serve as electroceuticals to provide less invasive and non-chemical-based methods to facilitate in vivo reprogramming or in situ direct cell conversion. 3D microenvironments can provide biomimetic systems to direct cell fate, including stemness maintenance, organogenesis, and tissue development. 252 - 254 Of particular interest are biomaterials-mediated and self-assembled cellular 3D microenvironments that can serve as powerful platforms to modulate cell reprogramming and highlight the clinical potential of reprogramming technologies. 3D biomaterial-based microenvironments enable robust spatial and temporal control of biophysical and biochemical cues to modulate cell fate both in vitro and in vivo . It has been demonstrated that mouse and human iPSC reprogramming can be enhanced in 3D PEG-based hydrogels by optimizing parameters such as stiffness, degradability, and biochemical composition. 227 It was determined that cell conversion could be vastly improved using highly degradable gels with a stiffness of 600 Pa. This 3D approach resulted in a more than 3-fold increase in the reprogramming efficiency compared to 2D environments, as a result of accelerating the MET process and increasing epigenetic remodeling. In addition, electrospun fibrous scaffolds, which have been widely applied as biomimetic microenvironments for biomedical applications, can also promote in situ stem cell neuronal reprogramming, resulting in neural network formation and brain engraftment. 216 Furthermore, several studies have shown that reprogramming cells in 3D microenvironments can enhance cell phenotypic functions. Chitosan substrate-mediated gene delivery of a naked Foxd3 plasmid can reprogram human fibroblasts into neural crest stem-like cells that self-assemble into 3D cellular spheroids. Notably, the Foxd3-transfected spheroids demonstrated significantly greater functional rescue of the CNS in impaired zebrafish models compared to that with dispersed cells. 255 In addition, under Matrigel-embedded 3D culture conditions, the combined expression of Gata6, Cdx2 , Hnf4a, and Foxa3 directly converted mouse fibroblasts into fetal intestine-derived progenitor cells (FIPCs) and allowed for the formation of spherical organoids, which further developed into budding organoids. 74 Particularly, the derived intestinal organoids regenerated colonic tissue upon transplantation, demonstrating great promise for the treatment of intestinal diseases. Therefore, these organoids, which consist of 3D structures with a heterogeneous cell composition, may serve as biomimetic systems that can recapitulate specific gene and protein expression, tissue morphology, and physiological functions. In the same study, it was determined that the defined transcription factors also induced human FIPCs from human dermal fibroblasts and human umbilical vein endothelial cells. 74 However, in contrast to mouse iFIPCs, human iFIPCs did not generate budding organoids, suggesting additional factors or functional 3D niches may be needed. Organ-on-a-chip is a burgeoning technique and powerful tool for disease modeling, drug discovery, toxicology research, and regenerative medicine. 256 Organ-on-a-chip systems are manufactured with advanced micro/nanofabrication techniques based on microfluidics, materials, and tissue engineering, enabling the design of customized cellular and tissue microenvironments with precise perfusion, mechanical and structural control, electrical stimulation, less probability of contamination, and high-throughput analysis. 257 The engineered architectures can recreate partial or all functions of native tissues, such as biological tissue barriers, parenchymal tissue functions, and multiorgan interactions. In particular, diverse 3D organoids, including liver, 258 pancreatic islet, 259 lung, 260 and brain, 261 that have been developed within organon-a-chip systems through in situ iPSC differentiation have demonstrated complex tissue and organ-specific functions. Most recently, organ-on-a-chip has shown great potential and superiority in cell reprogramming through the confinement of the microenvironment. Through the manipulation of the concentration and the sequential delivery of reprogramming factors (mixed mRNAs encoding transcription factors) and small molecules, an automated microfluidic platform dramatically improved human iPSC reprogramming with a 50-fold increase in efficiency ( Figure 10 ). 262 , 263 In the same platform, the generated hiPSCs could be precisely differentiated into functional cardiomyocytes and hepatocytes. The ability to perform both cell reprogramming and differentiation of patient-specific cells within these devices hold great promise for personalized drug screening, bringing us closer to more personalized medicine. Biochemical cues such as extracellular matrix (ECM) components, small molecules, growth factors, and other signaling molecules regulate not only stem cell fate but also cell reprogramming. 254 ECM and its derivatives are commonly used as hydrogels or bioactive molecules immobilized onto 2D substrates and 3D scaffolds. One study demonstrated that fibrin and a fibrin–collagen composite were superior to Matrigel and collagen I in supporting indirect cardiac reprogramming. 264 It was determined that matrix identity and tractional forces played major roles in cell reprogramming, whereas matrix mechanics, matrix microstructure, and cell proliferation influenced the process to a lesser extent. 264 To interrogate the effect of ligand identity, ligand density, and substrate modulus on cell reprogramming, ECM-coated polyacrylamide substrates can be developed to decouple physicochemical features. 265 In addition, PEG hydrogels can be modified with specific bioactive molecules to modulate specific cellular responses while preventing unwanted adsorption of proteins. PEG hydrogels modified with high concentrations of laminin and RGD peptide nearly doubled the efficiency of induced cardiomyocyte-like cells from fibroblasts. 265 Therefore, cellular reprogramming can be enhanced using customized engineered materials with immobilized biosignals. Small molecules and growth factors have been widely explored to enhance or induce either iPSC reprogramming 114 , 266 or direct reprogramming into various cell types, including neurons, 67 cardiomyocytes, 66 pancreatic β -like cells, 267 and hepatocytes. 268 Indeed, these small molecules and proteins can be mixed or conjugated, as ECM molecules, onto 2D cell culture substrates and 3D scaffolds to influence cell fate. Of note, such biochemical stimuli can be endowed and controlled by the defined delivery systems and can also be integrated with biophysical cues to maximize cell reprogramming, especially for in situ reprogramming.