Scaffold-Based Gene Therapeutics for Osteochondral Tissue Engineering

Significant progress in osteochondral tissue engineering has been made for biomaterials designed to deliver growth factors that promote tissue regeneration. However, due to diffusion characteristics of hydrogels, the accurate delivery of signaling molecules remains a challenge. In comparison to the direct delivery of growth factors, gene therapy can overcome these challenges by allowing the simultaneous delivery of growth factors and transcription factors, thereby enhancing the multifactorial processes of tissue formation. Scaffold-based gene therapy provides a promising approach for tissue engineering through transfecting cells to enhance the sustained expression of the protein of interest or through silencing target genes associated with bone and joint disease. Reports of the efficacy of gene therapy to regenerate bone/cartilage tissue regeneration are widespread, but reviews on osteochondral tissue engineering using scaffold-based gene therapy are sparse. Herein, we review the recent advances in gene therapy with a focus on tissue engineering scaffolds for osteochondral regeneration.


INTRODUCTION
Articular osteochondral injury is a common and frequently occurring disease in orthopedics, mainly caused by accidental trauma, sports injury or arthritis. Mature articular cartilage has a very weak ability to resist injury and disease, and has limited self-repair ability. After the articular cartilage is damaged, it cannot be effectively repaired, eventually leading to the occurrence of osteoarthritis (OA). It is expected that by 2030, OA will be the most common chronic degenerative joint disease among aging populations (Thomas et al., 2014;Tsezou, 2014). OA patients often suffer from severe pain and limited mobility. OA is also considered the leading cause of disability in the general population. The regeneration of articular cartilage that lacks self-healing ability is a major challenge in clinical treatment and clinically available methods fail to meet long-term effective regeneration requirements. This has caused concern in the field of osteochondral tissue engineering in which new tissues can be engineered to promote joint regeneration and prevent the onset of OA . One promising approach is the treatment of genes delivered by tissue engineering scaffolds. By transfecting specific gene sequences into seed cells, overexpressing or silencing the original gene, the biological function of the cells could be regulated to obtain the desired effect. Gene therapy combined with tissue engineering scaffolds provides a more precise, controlled, and sustained release of therapeutic factors compared to traditional methods of delivering growth factors directly ( Figure 1). This review focuses on recent advances in gene therapy in the field of scaffold-based osteochondral tissue engineering. In terms of miRNAs, we focus on recent research progress related to OA in the hope that miRNA can be used in the future gene therapy approaches combined with scaffoldbased osteochondral tissue engineering.

OSTEOCHONDRAL TISSUE ENGINEERING
Tissue engineering uses bionic scaffold to simulate the cell growth microenvironment and combines the body's selfhealing ability to guide tissue regeneration in damaged or defective tissue sites. The cell microenvironment of tissue engineering bionics can induce cartilage or the osteogenic differentiation of stem cells, promoting their proliferation and migration, leading to endogenous osteochondral regeneration . Osteochondral tissue engineering has evolved to enhance cell proliferation, differentiation, migration, and survival by transmitting growth factors and signaling molecules. These ligands combined with cell surface receptors of mesenchymal stem cells or mesenchymal progenitors, activate signaling pathways that promote osteochondral regeneration. However, the short half-life of recombinant proteins, such as BMPs, low bioactivity and high preparation costs lead to the exploration of new methods to deliver bioactive osteochondral regenerative compounds (Shi et al., 2014). A promising area is the use of scaffold-based gene therapy. By introducing specific gene sequences into cells, it is possible to modify or replace existing genes and regulate their epigenetic functions to achieve a desired purpose (Ginn et al., 2018). Gene-activated scaffolds provide a continuously controlled method of nucleic acid therapy to achieve a more efficient and safe release of biological agents.

Scaffold Biomaterials
To promote tissue regeneration, the osteochondral scaffold must be biocompatible, have a suitable rate of degradation, and possess a porous structure . To-date, osteochondral tissue engineering biomaterials include natural polymers, synthetic polymers, metals, and inorganic materials. Natural m a t e r i a l s a r e d e r i v e d f r o m a n i m a l s , p l a n t s , a n d microorganisms and can be classified into protein, polysaccharide, polyester, and polyamide based polymers according to their chemical composition (Nooeaid et al., 2012;Liu et al., 2018). The natural biological function and ability to promote cell adhesion and proliferation are unique advantages of natural polymer materials such as collagen, gelatin, and chitosan (Nooeaid et al., 2012;Kowalczewski and Saul, 2018). The variability and low mechanical strength of different batches of natural biomaterials lead to inevitable defects (Hsu et al., 2010). In contrast, mechanical properties can be carefully controlled through structural and surface modifications using synthetic polymers (Shimomura et al., 2014). However, because of its inherent hydrophobicity and lack of binding sites, their cell adhesion ability is relatively poor (Sarasam et al., 2006;Antonova et al., 2016). Another important consideration when designing osteochondral scaffolds is that the rate of degradation of biomaterials should match the rate of tissue repair. There are significant differences in the rates at which enzymes degrade natural polymers at different transplant sites in vivo, depending on the activity and concentration of the enzyme under different conditions. Conversely, hydrolytically degradable synthetic polymers show minor differences between sites or patients compared to enzymatically degradable polymers. However, the by-products of degradation are toxic (Zhang et al., 2014). Bioceramics, such as calcium phosphate, are characterized by their excellent osteoinductivity. Common types of bone calcium scaffolds are hydroxyapatite, tricalcium phosphate, biphasic calcium phosphate, and multiphase bioglass. By changing the composition of Ca 3 (PO 4 ) 2 ceramics, the stability and mechanical properties of the materials can be modified (Lima et al., 2019).
However, separate scaffold biomaterials are not effective in promoting osteochondral tissue repair. To increase the number of cells and the chondrogenesis/osteogenesis of MSCs, an array of cellular factors can be applied to the scaffold to promote and maintain the production of cartilage ECM.

Biochemical Factors
Bone morphogenetic proteins (BMPs) (Reyes et al., 2014), fibroblast-growth-factor 2 (FGF-2) (Yi et al., 2012), transcriptional SOX proteins (Cao et al., 2011), nel-like molecule-1 (Nell-1) (Zhang et al., 2016b;Wang et al., 2017), and IGF-1 and IGF-2 promote cartilage formation (Wang et al., 2009) and osteogenic differentiation. In addition, as angiogenic factors, Platelet derived growth factors (PDGF), vascular endothelial growth factor (VEGF), and early growth response gene 1 (EGR-1) promote bone repair (Franses et al., 2010;Press et al., 2015;Sheng et al., 2018). As anti-angiogenic factor, chondromodulin 1 (CHM-1) not only stimulates chondrogenesis but also inhibits chondrocyte hypertrophy and endochondral ossification (Klinger et al., 2011). A common route of administration for general growth factors is intravenous injection. However, the growth factor in the blood has a short half-life. By adjusting the physicochemical properties of the scaffold to slow release of growth factors, the drawbacks of direct administration can be avoided. Although 3D scaffolds can function as sustained-release growth factors, their ability to promote perivascular tissue healing and stem cells (SCs) regeneration is limited by their localization. Scaffold-based gene therapy provides a promising approach for tissue engineering through transfecting specific nucleic acids into cells to enhance the sustained expression of the growth factors of interest or through silencing target genes associated with bone and joint disease

MicroRNAs
Cartilage MicroRNAs (MiRNAs) are~22 nucleotide single stranded RNAs that regulate post-transcriptional gene expression. MiRNA induces degradation of the target mRNA by binding to the 3′untranslated region (UTR) complementary sequence on messenger RNA (mRNA), inhibiting translation, thereby suppressing corresponding protein production. Unlike small interfering RNA (siRNA), which regulates only one specific target, miRNA can regulate multiple targets. A single miRNA can regulate different targets in multiple signaling pathways, so it is more advantageous than other biomolecules in terms of functional effects. MiRNA expression profiles are significantly different during the development of articular cartilage, chondrocyte differentiation, and MSC chondrogenesis. Chondrocyte miRNA profiles differ from normal chondrocytes during their degeneration during osteoarthritis (OA). These miRNAs involved in chondrocyte differentiation or degeneration may be used in bioscaffolds in future studies to participate in the regeneration of cartilage tissue. The following is a summary of their latest.
Through its ability to target FUT1, microRNA-149-5p promotes the proliferation and survival of chondrocytes, thus preventing OA. It has also been found to be downregulated in patients with OA, leading to degenerative cartilage and disturbing homeostasis. Carriers have been employed to deliver miRNA-149-5p to MSCs to promote chondrogenesis (Celik et al., 2019). MiR-218 is highly expressed early in cartilage formation, but is stopped in synaptic-derived mesenchymal stem cells (SDSC) at the maturation stage of cartilage differentiation and miR-218 may directly regulates 15hydroxyprostaglandin dehydrogenase expression in SDSCs (Chen et al., 2019c). MiR-320c was decreased in the later stages of chondrogenesis of adipose-derived stem cells (hADSCs) and OA chondrocytes. It inhibits degeneration of OA chondrocytes by directly targeting beta-catenin and inhibiting Wnt signaling (Hu et al., 2019). MiR-92a-3p expression was increased in MSC chondrogenic exosomes and significantly decreased in OA chondrocytes exosomes. MiR-92a-3p may be involved in regulating cartilage development by targeting WNT5A (Mao et al., 2018b). Conversely, miR-182-5p plays a negative role in BM-MSC chondrogenesis by downregulating parathyroid hormone-like hormone (PTHLH) (Bai et al., 2019). Table 1 summarizes the miRNAs, related to cartilage development, that have been studied in recent years.
disadvantages, but in general, the transfection efficiency of current viral vectors is still higher than that of non-viral vectors.

Use of Adenoviruses
Adenoviral transgenic efficiency is typically close to 100% in vitro. Adenoviruses can transduce different human tissue cells, dividing and non-dividing. The production of high titer adenoviral vectors is simple and no integration into the genomes of human cells occur. As such, adenoviral vectors have been increasingly used in clinical trials of gene therapy and have become the most promising viral vectors, second only to retroviral vectors. In a recent study, it is found that the use of Adenoviral-BMP-2/basic fibroblast growth factor (bFGF)modified BMMSCs combined with demineralized bone matrix promote bone formation and angiogenesis, successfully repairing canine femoral head necrosis (ONFH) (Peng and Wang, 2017). However, the biggest challenge to the effectiveness of adenoviral approaches are the immune response.

Baculovirus Approaches
Baculoviruses show no pathogenicity toward humans and can be used under normal biosafety level 2 conditions. Baculoviruses, like adenoviral, induce both dividing and non-dividing cells. In some recent studies, baculoviruses has been used. Lo and colleagues employed Cre/loxP-based baculovirus vectors in adipose-SCs to enhance bone healing (Lo et al., 2017). Fu and coworkers highlighted the ability of baculoviruses to induce osteogenesis through allogeneic-MSCs (Fu et al., 2015). Despite this promise, the transient expression profiles of baculoviruses limit their use. In an attempt to overcome this issue, Chen and coworkers developed baculoviruses hybridized with the miR-155 scaffolds and the sleeping beauty transposon to sustainably inhibit transgene expression for extended time periods (Chen et al., 2011).

Lentiviruses
The advantage of lentiviral vectors are the high levels of foreign gene integration into the host chromosome in cells typically difficult to transfect, including primary cell cultures. Lentiviral vectors can be combined with chondroitin sulfate-hyaluronic acidsilk fibrin composite scaffolds and applied to bone-ligament connections to promote tissue engineering (Sun et al., 2014). In addition, Brunger et al. developed an independent bioactive scaffold that is capable of inducing stem cell differentiation and cartilage ECM formation using lentiviruses (Brunger et al., 2014). Despite the great progress in the study of lentiviral vector, it is still far from clinical application. First, the titer of recombinant virus is still not up to the level of in vivo application. Second, due to the complex biological properties of HIV, it is difficult to establish a stable HIV vector like the commonly used mouse retroviral vector, and the established packaging cells are not ideal.

Non-Viral Gene Delivery Vectors
Commercialized cationic lipids such as Lipofectamine 2000, Lipofectamine 3000, Lipofectamine RNAiMAX, and SiPORT NeoFx are widely used in biomaterial-based gene therapy. In recent studies, using lipofectamine 2000, Anti-miR-221 was transfected into adipose-MSCs which were seeded into synthetic nHA/PCL scaffolds. The results indicate that this method provides an effective way to promote osteogenesis of AT-MSCs (Hoseinzadeh et al., 2016). Macmillan et al. combined lipofectamine-complexed plasmids encoding BMP-2 and TGF-b1 with HA microparticles for delivery to the MSCs of three healthy pig donors. This study provides a promising approach to gene therapy that regulates stem cell growth and development to treat bone defects (McMillan et al., 2018). Although the toxicity of liposomes are well-known, more efficient transfection methods to replace them have not emerged. Recently, to enhance the interaction between cells and nucleic acids, Raftery et al. developed a new cell penetrating peptide, GET, combined with a variety of collagen scaffolds, which showed good regeneration potential. GET is suitable for all three germ layer cell transfections with efficiencies comparable to Lipofectamine 3000 and minimal cytotoxicity. These findings suggest that GET can be combined with scaffold delivery systems, to provide new solutions to a variety of tissue engineering regenerative indications (Raftery et al., 2019).

GENE THERAPY IN SCAFFOLD BASED OSTEOCHONDRAL TISSUE REPAIR
Gene therapy for osteochondral tissue repair is divided into two phases: one to locate the gene to the target area directly, either through encapsulation onto a scaffold, or through a specific gene vector (in vivo). Alternatively, the target gene is loaded into the cells by the vectors in vitro, and genetically modified cells are administered to the target lesion area, with or without a scaffold (ex vivo). However, the main obstacle to the treatment of focal defects with non-scaffolds is that the genetically modified cells or gene vectors with intra-articular injections are diluted by the joint fluid and fail to reach the target lesion area. To avoid this drawback, a promising approach is to deliver modified cells or gene vectors using different types of scaffolds. When the scaffold is degraded, the contents are slowly released to the target area. Gene therapy combined with scaffolds increases the efficiency and duration of transfected genes, forming an efficient system to promote osteochondral regeneration. We herein summarize and discuss these gene therapy-binding scaffolds discovered from 2006 to 2019 in the contest of seeding cell types (Figure 2).

Gene Modified BMSCs
MSCs are the most widely studied due to their high availability and proliferative/differentiation ability. The microenvironment typically dictates the fate of MSCs. BMSCs are more commonly employed than those derived from adipose tissue (AMSCs), particularly for osteochondral therapy. In one study, BMSCs were transfected with hIGF-1 cDNA and mixed with calcium alginate gels for transplantation into 6 mm osteochondral defects and were found to improve the repair (Leng et al., 2012).
In view of the role of TGF-b in promoting cartilage repair, in addition to the inhibition of inflammatory and immune responses, pcDNA-TGF-b gene-modified BMSCs were seeded onto biodegradable poly-L-lysine coated polylactic acid (PLA) scaffolds which were transplanted into New Zealand rabbit articular cartilage full-thickness defects. In vitro, after 2 weeks of cell seeding, the cartilage matrix formed and filled with the attachment holes of the scaffold. In vivo, 24 weeks after transplantation, the hyaline cartilage repaired the cartilage defect area, trabecular bone and dense bone repair in the subchondral area and the quality of the regenerated tissue improved over time (Guo et al., 2006). Similarly, BMSCs were transduced with pDNA-TGF-b1 and loaded into PLGA/fibrin gel hybrids which were implanted into rabbit articular cartilage models, resulting in the regeneration of new cartilage tissue with similar thickness, cell arrangements, color, and abundant glycosaminoglycans to normal cartilage . Moreover, TGF-b1-gene engineered rat BMSCs induced cartilage regeneration in rats (He et al., 2012), and their transfer onto PGA scaffolds using adenoviral approaches induced chondrogenic differentiation both in vitro and in vivo (Xia et al., 2009). Robust collagen II staining was observed in adenovirus-mediated-BMP-2 and TGF-b3 infected cells. DBM compounds with BMP-2 and TGF-b3 infected BMSC scaffolds showed high biocompatibility and the capacity to regeneration cartilage in pig models .
SOX9 is a transcription factor of the SOX (Sry-type HMG box) protein family that promotes cartilage formation and the phenotypes of chondrocytes. Adenoviral vectors have been used to transfect SOX9 into rabbit BMSCs which effectively induced their differentiation into chondrocytes on PGA scaffolds and improved the repair of cartilage defects (Cao et al., 2011). More recently, Venkatesan et al. designed 3D fibrin-polyurethane scaffolds in a hydrodynamic environment that provided a favorable growth environment for rAAV-infected SOX9modified hBMCs and promoted their differentiation into chondrocytes. Interestingly, the expression of SOX9 lasted 21 days, the longest time point evaluated (Venkatesan et al., 2018).
Cartilage engineering can also be mediated through gene activation matrices. Rowland et al. engineered lentiviral particles expressing a doxycycline-inducible IL-1 receptor antagonist (IL-1Ra) on a cartilage-derived matrix to prevent IL-1 mediated inflammation. Similar scaffolds have been employed for site directed chondrogenic and osteogenic differentiation using BMSC populations that overexpress either chondrogenic, BMP2, or TGF-b3 transgenes. The ability to regulate IL-1Ra expression afforded protection to the cartilagematrix in the presence of IL-1, leading to enhanced bone production and cartilage formation. When inflammation is absent, cartilage-derived matrix hemispheres expressing TGF-b3 and BMP-2 were also fused to the bilayers of osteochondral constructs to promote healing (Rowland et al., 2018). Yang and coworkers also transfected BMSCs with adenoviruses expressing C-type natriuretic peptides and seeded the cells onto silk/ chitosan scaffolds to promote chondrogenesis in rat cartilage defect models .
Despite advances in the development of osteochondral repair scaffolds, their combination with miRNAs remains in the early stages. MiR-221 can induce BMSCs toward chondrogenesis in the absence of TGF-b and could repair osteochondral defects following its subcutaneous implantation into mouse models, promoting Collagen type II positive tissue expression that was negative for collagen type X (a well characterized marker of hypertrophy). The potential of hMSCs silenced for miR-221 to suppress collagen type X represents an exciting development with clear therapeutic potential for cartilage repair in the clinic (Lolli et al., 2016).

Gene Modified ADSCs
It is now well accepted that ADSCs have clinical utility. An advantage is their ability to be collected via lipoaspiration, a noninvasive harvesting process. Lu and coworkers developed baculoviruses for FLPo/Frt expression of genetically engineered rabbit ADSCs. The cells were transfected with TGF-b3/BMP-6 and added to PLGA-GCH scaffolds for implantation to cartilage defects in weight-bearing areas, promoting regeneration. The designed neo-cartilages had defined cartilage-specific-structures in the absence of degeneration or hypertrophy (Lu et al., 2014). In other studies, the inguinal fat of rats were transduced with SOX via retroviral approaches and ADSCs were collected and seeded into fibrin gels and implanted onto defects in the femur patellar groove. These approaches significantly increased type II collagen expression, GAG levels, and improved cartilage healing (Lee and Im, 2012). Upon seeding the ADSCs into large PCLscaffolds immobilized with Dox-inducible lentiviruses expressing IL-1Ra, controlled tissue growth and biomimetic cartilage properties were maintained (Moutos et al., 2016).

Gene Modified Chondrocytes
Isolated cartilage cells can be obtained through enzymatic digestion and can embed into cartilage lacuna, preventing immune cell invasion and organ rejection. However, the cells dedifferentiate overtime and their propensity for cartilage production becomes impaired, limiting their use in clinical application. The use of 3D cultures can mimic the microenvironment of the extracellular matrix permitting the maintenance of phenotypic stability. In this regard, neonatal male foals chondrocytes transduced with IGF-1-adenoviruses and embedded into fibrinogen were implanted into equine defects and conferred high levels of IGF-1 expression and cartilage healing (Goodrich et al., 2007). Griffin and colleagues used a comparable approach with rAAV5 and implanted the carriers into equine femurs, also showing improved graft healing (Griffin et al., 2016). FGF-2 and IGF-I plasmid vectors have also been delivered into Lapine articular chondrocytes. The cells were encapsulated into alginate scaffolds and transplanted onto rabbit knee joint defects for a period of three weeks, in which enhanced IGF-I/FGF-2 levels improved the defects with no adverse effects to the synovial membrane, highlighting the utility of these approaches to promote cartilage repair (Orth et al., 2011).
FGF-2 is mitogenic in articular chondrocytes and when transfected into articular chondrocytes and encapsulated in alginate scaffolds, FGF-2 expression was maintained for over 21 days and improved cartilage defects in the knee joints of rabbits. No adverse effects were again evident in the synovial membrane following histological assessments but type II collagen expression was enhanced (Kaul et al., 2006).
Mechanical movements activate integrin b1-signaling and enhance the proliferative capacity of chondrocytes, increasing matrix synthesis. Liang and co-workers seeded integrin b1transfected chondrocytes onto PLGA scaffolds which produced higher levels of GAG and type II collagen after lentiviral-integrin b1 transfection compared to mechanically stressed sham controls. The opposing phenotype was observed in the cells silenced for integrin b1, suggesting that in addition to mechanical stimulation, the overexpression of integrin b1 enhances cartilage regeneration (Liang et al., 2015) (Table 3).

CONCLUSION AND FUTURE DIRECTIONS
In summary, osteochondral defects are not a single cartilage or bone injury, but involve complex multi-structural components. The healing of these components is challenging. To-date, there is no technology that can form a natural cartilage structures in the joints. Osteochondral tissue engineering shows good potential for osteochondral repair and OA treatment, but several problems remain. For example, at the seed cell level, chondrocytes have poor availability and dedifferentiation properties. Unacceptable outcomes such as chondrocyte hypertrophy and endochondral ossification are often accompanied by an inability to control the differentiation of chondrogenic SCs. Also, due to its unique layered structure, osteochondral tissue theoretically requires a multi-phase structure to simulate the native layered structure, but this is difficult to achieve. Recent studies have shown that a combination of gene vectors, genes, seed cells, and scaffolds are more likely to obtain hyaline cartilage, with the combined changes between them primarily based on lesion size, location, and structure.
Genes have been transfected into MSCs or chondrocytes to improve their phenotypic properties. In general, cartilage gene therapy enables seed cells to continuously encode growth factors, transcription factors, or anti-inflammatory cytokines, thereby inducing cartilage differentiation and inhibiting the progression of inflammatory diseases (Figure 2). Studies have shown that multiple combinations of genes encoding growth factors, transcription factors, or anti-adverse response cytokines are more advantageous than single genes for improving healing and reducing adverse effects. To minimize hypertrophy, ossification, and host immune responses, complex gene delivery vectors must be designed to increase safety and more sustained gene protein release. miRNAs regulate chondrogenesis and arthritis. The expression of a specific miRNA mimetic or miRNA inhibitor permits the manipulation of the expression profiles of the cellular miRNAs and their epigenetic features. On this basis, combined with 3D biological scaffold printing technology, it is more conducive to accurately control cell differentiation and optimize the biochemical and biomechanical properties of regenerated tissues. However, the use of 3D delivery systems to miRNAactivated scaffolds is in its infancy. Moreover, in terms of scaffolds, 3D multiphase structural scaffolds are complex, and not conducive to the control of each phase, including degradation rates and shear forces. Therefore, the two-phase scaffold divided into a cartilage phase and a bone phase is simpler than multi-phase scaffolds and ideal for osteochondral scaffolds (Seo et al., 2014). We propose that to make full use of the integrated fusion bilayer scaffold, each genetically modified cell line (overexpression or knockout of miRNA) can edit specific signaling molecules that facilitate tissue regeneration in each layer.