BMP gene delivery for skeletal tissue regeneration

Abstract

Musculoskeletal disorders are common and can be associated with significant morbidity and reduced quality of life. Current treatments for major bone loss or cartilage defects are insufficient. Bone morphogenetic proteins (BMPs) are key players in the recruitment and regeneration of damaged musculoskeletal tissues, and attempts have been made to introduce the protein to fracture sites with limited success. In the last 20 years we have seen a substantial progress in the development of various BMP gene delivery platforms for several conditions. In this review we cover the progress made using several techniques for BMP gene delivery for bone as well as cartilage regeneration, with focus on recent advances in the field of skeletal tissue engineering. Some methods have shown success in large animal models, and with the global trend of introducing gene therapies into the clinical setting, it seems that the day in which BMP gene therapy will be viable for clinical use is near.

1. Introduction

Musculoskeletal disorders currently affect more than 10 million persons in the USA. This number is expected to rise significantly as the world population ages [1]. These disorders have been historically underappreciated in both clinical care and research, because they are not considered fatal. Nevertheless, musculoskeletal disorders are associated with considerable morbidity.

Fractures are undervalued because bone tissue can heal naturally as long as the fracture is stabilized properly [2–4]. Several conditions such as trauma, bone tumor resection, or arthritis, however, can lead to large bone defects that may not heal well. The prevalence of bone nonunion among people with fractures is currently estimated to be 4.9%; as the world population ages, this rate may increase [5]. Despite early detection and prompt treatment, the 30-day and 1-year mortality rates associated with hip fractures are 7.7% and 26%, respectively [6]. Bone fractures occur frequently, are costly, and often produce devastating physical and emotional effects. Thus, improved methods to regenerate bone would be useful in a wide variety of scenarios and could decrease mortality rates. Cartilage defects are also associated with substantial morbidity, because they have a very limited capacity for self-regeneration and, if left untreated, will progress to osteoarthritis [7].

Currently, the gold-standard treatment for bone loss and other substantial tissue damage is autologous grafting (with the aid of autografts). However, autografts are not always available, and their harvest often leads to prolonged postoperative pain and substantial donor site morbidity [8]. Bone allografts are readily available from tissue banks and, therefore, could be an alternative to autografts. Unfortunately, allografts have a very low osteogenic potential; this leads to poor graft-host integration and results in numerous bone regeneration failures [9].

One way to treat nonunion bone fractures in lower limbs is the Ilizarov technique, also referred to as bone transport or distraction osteogenesis. Using this technique, an external circular modular fixator is affixed to the broken bone via heavy-gauge wires. The fixator allows for partial weight bearing while applying tension to the fractured bone, thus inducing gradual bone regeneration. Disadvantages of this technique include pain, poor patient compliance, the inconvenience of wearing the frame, and the risk of inducing bone malalignment; in addition, the Ilizarov technique is a complicated procedure that requires a learning curve on the part of the surgeon [10,11].

Bone morphogenetic proteins (BMPs) are a group of transforming growth factor-β (TGF-β) proteins that play a crucial role in bone remodeling and repair [12]. In recent years, recombinant human BMP proteins have been introduced into the clinical setting for a variety of indications including long-bone fracture repair. Local administration of BMP2 and BMP7 in patients with tibial nonunion fractures resulted in increased healing rates [13,14]. However, the sales of BMP7 product were discontinued in 2014. The remaining product, BMP2, is costly and requires supraphysiologic concentrations to induce bone healing. Unfortunately, it has been associated with a high incidence of side effects such as inflammation [15], inhibition of bone formation [16], bone cysts, and ectopic bone formation [16]; and in some cases, related to spine surgery, it has led to neurological impairment [17]. These disadvantages have spurred many researchers to develop new strategies to overcome lack of bone regeneration at sites of bone loss.

Gene therapy is one alternative method of introducing therapeutic proteins into the body. The main advantages of gene therapy include the ability to control the duration of gene expression and to target specific organs where the gene can be expressed in a physiologic manner. BMP gene therapy can provide a sustained secretion of the protein in a localized manner, i.e. at a fracture site, and in some cases, by specific cells that may enhance the therapeutic effect. By that, transgene expression mimics the natural process of bone healing, without the need for high doses of a growth factor. Studies have shown that nonviral BMP gene therapy can induce the secretion of picogram levels of the protein [18,19], compared to micrograms that are included in the commercial product. In addition, it has been shown that only a short term of BMP2 or 6 expression is required to obtain bone regeneration and fracture healing in animal models [19,20].

Both viral and nonviral methods of gene therapy are available. Viral vectors are well studied; they offer ease of use and high efficacy at transducing cells in vivo. More than 2000 viral gene therapy clinical trials have been completed worldwide to date [21]. However, this type of therapy is attended by safety concerns that currently limit its use in clinical practice, namely the use of viruses and small amounts of payload genes. Alternatively, several nonviral methods of gene delivery have been developed to overcome the disadvantages of using viral vectors. These methods include use of liposomes, polycations, electroporation, sonoporation, and gene-activated matrices (GAMs). Unfortunately, these methods are currently limited by relatively low efficacy and require large gene doses to reach the efficacy of virus use [21]. Gene delivery can be done in vivo, directly to the target tissue, or it could be performed ex vivo upon a selected population of cells that will later be delivered to the desired location. Table 1 describes the pros and cons of the various vectors and gene delivery approaches.

Table 1.

Pros and cons of gene delivery methods.

Method	Pros	Cons
In vivo gene delivery	Straightforward	Difficult to target specific cells or tissues Risk of unwanted biodistribution Less control over efficiency of expression
Ex vivo gene delivery	Better control of targetted cells and expression level	Complex protocols of cell isolation and expansion might lead to a costly product and a difficult regulatory pathway.
Viral vectors	High efficiency of gene delivery	Risk of an immune response Risk of tumorigenicity due to insertional mutagenesis Stable integration might lead to an excess of protein secretion Production might be costly and complex
Non-viral vectors	Considered safer compared to viral vectors. Simpler production process Easier manipulation	Low efficiency of gene delivery

BMP-encoding genes are the most studied genes for skeletal tissue regeneration. Among the 14 different BMP genes, BMP2, BMP6, and BMP9 were found to be the most potent inducers of osteogenic differentiation, with some evidence showing that BMP6 and BMP9 are more potent than BMP2 [18,22]. Since then, many studies have evaluated the efficacy of these genes in bone repair when delivered using various methods. In this review, we will present recent advances in BMP gene therapy for bone regeneration.

2. In vivo gene delivery (Table 2)

Table 2.

In vivo gene delivery for bone regeneration.

Method	Vector	Carrier	Gene	Species	Model	Reference
Viral vectors	Adenovirus		BMP2	Rat	Femoral bone defect	[23,42]
	Adenovirus		BMP2	Rabbit	Femoral bone defect	[24]
	Adenovirus		BMP2	Sheep	Iliac crest and tibial defects	[25]
	Adenovirus	Silk scaffold	BMP7	Mouse	Calvaria bone defect	[29]
	AAV	Bone allograft	BMP2	Mouse	Calvaria bone defect	[27]
	AAV	Bone allograft	BMP2	Mouse	Femoral bone defect	[28]
Nonviral vectors	GAM	Collagen-hydroxyapatite	VEGF	Human	Mandible nonunion	[33]
	GAM	Collagen	BMP2 (mRNA)	Mouse	Calvaria bone defect	[38]
	GAM	Fibrin gel	BMP2 (mRNA)	Rat	Femoral bone defect	[39]
	Electroporation		BMP9	Mouse	Radius bone defect	[41]
	Electroporation		BMP2/7	Rat	Alveolar bone regeneration	[43]
	Electroporation		BMP2/VEGF	Rabbit	Mandibular distraction osteogenesis	[44]
	Sonoporation		BMP9	Mouse	Ectopic bone	[47]
	Sonoporation		BMP2/7	Rat	Femoral bone defect	[48]
	Sonoporation		BMP6	Pig	Tibial bone defect	[19]
	Liposome		BMP2	Pig	Peri-implant bone defects	[53]

2.1. Viral vectors

A wide variety of gene delivery strategies exist. Use of viral vectors is the most widely researched technique due to the ease of viral vector production and use, and their high efficacy rates. Commonly used viral vectors include adenoviruses, adeno-associated viruses (AAVs), and lentiviruses. Early studies showed that adenoviral vectors encoding BMP2 can be used to treat femoral defects in rodents [23]. Several large animal studies showed promise in bone healing using BMP2-encoding adenoviral vectors. Baltzer et al. demonstrated that local injection of this vector can lead to complete segmental defect repair in the femur bones of rabbits [24]. The same approach had success in healing iliac crest and tibia bone defects in sheep [25]. Due to several safety concerns associated with adenoviruses, AAVs are considered safer because they have not been associated with human diseases [26]. Compared to other vectors, AAVs lead to long-term gene expression due to genomic formation of episomal concatemers in the nucleus while maintaining low immunogenicity. More recently, adenoviruses and AAV vectors have been used together with scaffolds to enhance bone formation. AAV-BMP2-coated allografts were shown to be superior to uncoated allografts in craniofacial bone regeneration [27] and in femoral fracture repair in mice [28]. Zhang et al. used porous silk scaffolds containing a BMP7-encoding adenoviral vector to regenerate calvarial defects in mice within 4 weeks [29].

2.2. Nonviral vectors

2.2.1. Gene-activated matrices

An alternative approach to viral vectors is nonviral gene delivery.Many different materials, including cationic polymers, lipids, peptides, and calcium phosphate can be used to carry genetic information to target cells in vivo [30]. One of the main benefits of gene delivery in skeletal tissue engineering is the ability to deliver genes in a localized fashion. Since skeletal defects are often localized, such an approach is especially attractive. Biodegradable materials are good candidates for such an approach, because they allow local delivery in a sustained manner to nearby cells [31]. Gene-activated matrices (GAMs) are scaffolds containing plasmid DNA, which allow for a controlled slow release of genetic material to the environment [32]. Some of these scaffolds have structural properties that provide mechanical support during the regenerative process. In addition, this method allows for sustained yet transient expression of inserted genes, as they do not integrate into the cell genome, which is ideal for bone fractures. The method has already been tested in a human patient, in whom a collagen-hydroxyapatite scaffold and plasmid DNA encoding for vascular endothelial growth factor (VEGF) was implanted in a mandibular nonunion fracture [33]. No adverse effects were observed 1 year following implantation, and bone regeneration was observed. Yet, at its current stage, this method has relatively low efficacy in gene delivery and requires high doses of plasmid DNA to induce significant effects [34]. Various polymers have been used to improve upon gene delivery of BMP plasmid DNA. Polyamidoamine dendrimers, spherical hyperbranched molecules with a dense central core, have shown particular promise in delivering BMP2 to mesenchymal stem cells (MSCs) in various models [35,36].

Given that angiogenesis is an important factor in bone regeneration, it is clear that bone regeneration can be further enhanced by delivering angiogenic factors. Curtin et al. used GAMs to deliver BMP2 and VEGF for bone regeneration in a calvarial defect in rats [37]. In that study, the authors compared the efficacy of collagen nano-hydroxyapatite and polyethyleneimine in delivering BMP2. Interestingly, while BMP2 expression was favorable when the polyethyleneimine vector was used, higher osteogenic activity was present when the collagen nano-hydroxyapatite vector was used, emphasizing the importance of vector choice.

Delivery of messenger RNA (mRNA) has also been evaluated for bone regeneration as an alternative to plasmid DNA. As mRNA delivery targets the cytoplasm, immediate expression of the encoding protein results even at low transfection rates [38]. For more stability and less immunogenicity, chemical modifications are required for this technique to become effective in tissue engineering. Balmayor et al. used biphasic calcium phosphate vectors and fibrin gel as BMP2 mRNA carriers and showed that mRNA can be used to regenerate femoral bone defects in rats [39].

2.2.2. Physical methods of cell membrane poration

Electroporation uses high-voltage pulses to form transient pores in the cell membrane, allowing free diffusion of genetic material into the host cell [40]. Kimelman-Bleich et al. used electroporation to deliver BMP9 DNA plasmid to endogenous cells residing within fracture sites in mice [41]. This method proved effective in repairing segmental bone fractures within 5 weeks. For such gene delivery methods to work, a viable endogenous cell population at the target site is required. The authors implanted a collagen scaffold and waited 10 days before gene delivery. This two-step method proved to be more efficient than delivering the gene soon after the surgical procedure. Betz et al. compared the efficacy of adenoviral BMP2 gene delivery to bone fracture sites at various time points: intraoperatively or at 1, 5, or 10 days after surgery [42]. Those authors showed that delaying gene delivery 5 or 10 days after surgery allows for better bone regeneration with improved mechanical properties. More recently, electroporation was used to deliver BMP2/7 plasmid DNA constructs to periodontal tissues to increase the rate of alveolar bone formation in rodents [43]. In addition, BMP2 and VEGF were also co-delivered to enhance osteogenesis using electroporation, in order to heal mandibular defects in rabbits [44].

An alternative to electroporation, sonoporation, has been investigated as a physical nonviral method of gene delivery for bone regeneration. Sonoporation uses ultrasonic waves to increase cell membrane permeability by forming transient nanosized pores in the membrane, subsequently enhancing uptake of drugs and nucleic acids [45,46]. Sheyn et al. showed that delivery of BMP9-encoding DNA using ultrasound with microbubbles induced ectopic bone formation in a mouse thigh [47]. Later, Feichtinger et al. used a similar approach to deliver BMP2 and BMP7-encoding DNA in a rodent femur fracture model [48]. Despite successful induction of ectopic bone formation, the authors were unable to induce complete fracture repair. More recently, ultrasound was successfully used to deliver a BMP6 gene and regenerate critical-size defects in tibia bone [19] and to enhance the osteointegration of anterior cruciate ligament grafts [49] in pigs. Yet, sonoporation has many different components that affect the efficacy of gene delivery, hampering its applicability in various models. Shapiro et al. optimized the efficiency of sonoporation by comparing different combinations of exerted acoustic power and plasmid DNA concentrations, treatment durations, and number of treatments [50].

2.2.3. Lipid-based vectors

Liposomes consist of concentric lipid bilayers and can encapsulate genetic material for gene delivery via endocytosis and release of DNA into the cell cytoplasm [51]. Park et al. compared liposomes and an adenoviral vector as aids to in vivo delivery of the BMP2 gene for bone regeneration [52]. Although the authors did not quantify the resultant bone formation, histological data showed similar results for both groups. The same researchers later applied liposomal BMP2 gene delivery, which proved to enhance bone regeneration around dental implants in pigs’ calvarial bone [53]. Several lipid-based formulations are commercially available (e.g., Lipofectamine™) and used mainly for in vitro gene delivery. Currently, clinical use of these lipid-based vectors is hindered due to their cytotoxicity, and thus safer and more efficient formulations are needed for clinical translation [54].

3. Ex vivo gene delivery

3.1. Bone regeneration (Table 3)

Table 3.

Ex vivo gene delivery for bone regeneration.

Vector	Cell	Gene	Species	Model	Reference
Adenovirus	Bone marrow-derived MSCs	BMP2	Mouse	Radius bone defect	[56]
Electroporation	Bone marrow-derived MSCs	BMP2 or BMP9	Mouse	Ectopic bone	[57]
Electroporation	Bone marrow and adipose-derived MSCs	BMP2 or BMP6	Mouse	Ectopic bone	[18]
Electroporation	Bone marrow-derived MSCs	BMP6	Pig	Vertebral bone defect	[58]
Electroporation	Adipose-derived MSCs	BMP6	Rat	Vertebral bone defect	[59]
Lentivirus	Umbilical cord blood -derived MSCs	BMP2	Mouse	Ectopic bone	[60]
Lentivirus	Bone marrow-derived MSCs (freshly - isolated or sub-cultured)	BMP2	Rat	Femoral bone defect	[65]
Adenovirus	Gingival fibroblasts	BMP2	Rat	Calvaria bone defect	[61]
Adenovirus	Periosteal cells	BMP2	Rabbit	Mandible bone defect	[62]
Baculovirus	Adipose-derived MSCs	BMP-2 and miR-148b	Mouse	Calvaria bone defect	[73]
Polyethylenimine (PEI)	Bone marrow-derived MSCs	miRNA-20a and siRNA against noggin	Rat	Calvaria bone defect	[78]
Baculovirus	Adipose-derived MSCs	BMP2 and CRISPR interference against noggin	Rat	Calvaria bone defect	[80]

Ex vivo gene delivery, or cell-mediated gene therapy, uses in vitro cell transfection followed by in vivo delivery of the therapeutic cells. This approach allows better control over the transfection process and the target cell population, and introduces cells with regenerative capacities directly into the injury site [55].

Local delivery of engineered human bone marrow-derived MSCs expressing BMP showed early promise in rodents [18,56,57] and was later adopted by Pelled et al., who achieved complete bone regeneration within 6 weeks using BMP6-expressing MSCs in a vertebral defect model in pigs [58]. Other studies showed similar successes in various models using BMP-engineered adipose-derived MSCs [59], umbilical cord-derived MSCs [60], gingival fibroblasts [61], and periosteal cells [62,63]. The question remains regarding the optimal cell source for bone regeneration. It was initially thought that MSCs would provide the most promising results, because they are able to directly participate in bone regeneration. However, Gugala et al. demonstrated no significant differences in osteoinduction between various cell sources that were BMP2 transduced ex vivo, concluding that cell mediated gene therapy for bone regeneration is independent of cell type [64].

In a recent attempt to reduce the production time of engineered cells for clinical use, researchers compared the regenerative capacity of freshly isolated human bone marrow cells to that of bone marrow cells expanded in culture after both cell types had been transduced with lentiviral-BMP2 vector [65]. Both types of transduced cells were implanted into critical-size femoral bone defects in rats. While application of both cell types led to enhanced osteogenesis, use of the culture-expanded cells produced higher fracture union rates, from which one can conclude that culture expansion is still an essential step in cell-based bone regeneration.

Cell-mediated tissue regeneration often requires the use of biodegradable scaffolds to carry the cells and retain them at the implantation site. Scaffolds also support the growth and differentiation of the loaded cells and could serve as a three-dimensional (3D) template to facilitate bone repair. Many factors, including porosity, mechanical properties, and biodegradability, affect the scaffold’s ability to repair bone [66]. Scaffolds seeded with MSCs were shown to improve bone regeneration better than acellular scaffolds [67]. Additional studies also showed that the chemical composition of the scaffold has a direct effect on the osteoinductive properties of seeded MSCs [68,69].

Several studies have found that BMP-engineered MSCs implanted in scaffolds improve the regenerative capacity of the constructs compared to naïve MSCs. In one study, encapsulated MSCs were transduced with BMP2 within a hydrogel scaffold by using advanced visible light-based projection stereolithography technology [70]. This construct was compared to encapsulated naïve MSCs provided with exogenous BMP2. The authors found that the BMP2-engineered cells displayed prolonged expression of BMP2 and promoted significantly more bone formation in vivo.

Cells can also be engineered to express miRNAs, a family of small non-coding RNAs that silence post-transcriptional expression, have recently attracted great attention in tissue engineering [71]. Zhang et al. showed that miR-20a can increase osteogenic differentiation in vitro through inhibition of peroxisome proliferator-activated receptor γ (PPARγ), which in turn leads to overexpression of BMP2 [72]. Liao et al. implanted human adipose-derived stem cells that had been co-transduced with BMP2 and miR-148b in rodent calvarial defects [73]. The authors showed that the combination of BMP2 and miR-148b significantly increases bone regeneration and leads to near-complete repair of calvarial defects within 12 weeks. A similar approach showed that downregulation of BMP antagonists, such as Noggin and Chordin, resulted in enhanced osteogenic potential of MSCs and bone regeneration in vivo [74–80]. Methods of downregulation or knockdown included the use of small hairpin RNA (shRNA) [76], small interfering RNA (siRNA) [74,75,77–79] and CRISPR interference (CRISPRi) [80]. The use of CRISPR technology could be potentially used to overexpress BMP genes. However, as discussed above, since the BMP gene only needs to be expressed for a short term, prolonged expression due to stable integration should be avoided, as it might lead to excessive secretion of the protein.

3.2. Non-osseous skeletal tissues (Table 4)

Table 4.

BMP gene delivery for non-osseous skeletal tissue regeneration.

Vector	Cell	Gene	Species	Model	Reference
Sonoporation		BMP6	Pig	Anterior cruciate ligament reconstruction	[49]
Adenovirus	Muscle or fat tissue biopsy	BMP2	Rabbit	Osteochondral defect	[83]
Baculovirus	Adipose-derived MSCs	TGF-β3/BMP-6	Rabbit	Articular full defect	[84]
Adenovirus	Tendon cells	BMP12	Chicken	Full thickness tendon tear	[88]
Lentivirus	Muscle-derived stem cells	BMP2	Rats	Articular cartilage repair	[89]

Ex vivo gene therapy for cartilage regeneration is especially attractive because it introduces an abundant cell source to the avascular cartilage space [81]. Chitosan-based scaffolds implanted with BMP6-transfected MSCs were shown to be effective in inducing in vitro chondrogenesis [82]. Evans et al. used BMP2-adenoviral vectors to transduce tissue autografts that were implanted into osteochondral defects [83].The authors showed improved healing with histological continuity and good fusion of activated autografts with adjacent healthy cartilage. Lu et al. implanted TGF-β3/BMP6-engineered adipose-derived stem cells into poly(lactic-co-glycolic) acid (PLGA) scaffolds, which were later implanted into full-thickness articular cartilage defects in rabbits [84]. The scaffolds were conjugated with gelatin, chondroitin-6-sulfate, and hyaluronic acid to augment chondrogenesis. The implantation resulted in complete regeneration of cartilage defects 12 weeks after implantation.

Gene delivery of BMP12, BMP13, and BMP14 has been studied as a potential method for tendon regeneration. These proteins were shown to promote elastin and collagen I expression during embryogenesis and to improve the biomechanical properties of treated tendons [85,86]. It was shown in vitro than BMP12 is an essential factor required for the differentiation of MSCs into tendon cells [87]. Lou et al. showed that BMP12-transduced tendon cells that were locally administered resulted in improved tendon regeneration, with a two-fold increase in tensile strength in a chicken model of tendon laceration [88].

It remains unclear, however, whether cell-mediated BMP gene therapy provides better results than use of controlled-release delivery vehicles. Gao et al. utilized human muscle-derived stem cells (hMDSCs) that had been transduced using a lentiviral-BMP2 vector in vitro [89]. The hMDSC-BMP2 cells were then injected directly into monoiodoacetate-induced osteoarthritic knees in a rat model. This method was then compared to delivery of BMP2 and hMDSCs via coacervate sustain release technology. The authors concluded that use of coacervate can achieve cartilage repair similar to that produced using virally transduced cells. Interestingly, the endogenous cells recruited by BMP2 signaling were more responsible for bone regeneration than the transplanted cells. Indeed, Loozen et al. seeded scaffolds with MSCs or fibroblasts transfected with BMP2, BMP6, or BMP7 [90]. The seeded scaffolds were then implanted ectopically in rodents. The authors found that constructs containing BMP-transfected MSCs and those containing BMP-transfected fibroblasts attained similar bone formation, which was significantly higher than that attained by naïve MSCs, emphasizing the importance of the paracrine activity of transfected cells.

4. Conclusions

Gene therapy for skeletal tissue regeneration has shown great promise in animal models during the last 20 years, mainly in bone regeneration and to a lesser extent in cartilage and tendon repair. In this review, we covered recent advances in the use of BMP-associated gene delivery. Since BMP2 and BMP7, as recombinant proteins, are approved by the U.S. Food and Drug Administration, translation of gene delivery systems, encoding for these genes, into clinical practice would have been expected by now. Clearly more efforts need to be invested in order to achieve successful translation of BMP-based gene therapy. First, certain safety concerns need to be addressed. Specifically, biodistribution analysis needs to be done in relevant large animal models in order to rule out off-target gene expression that might lead to unwarranted bone formation. Previous studies showed an immune response to adenovirus and AAV vectors encoding for BMP6 [91]. Such a response should be investigated for other viral vectors or naked DNA. Finally, a large number of publications indicated that BMP2 could enhance the activity of various tumor types, albeit not induce cancer ne novo [92]. This point should be taken into consideration in cases where tissue regeneration is required following tumor resection, due to the hazard of remaining cancerous cells.

Additional hurdles towards clinical translation may be attributed to clinical needs. To date, bone regeneration is mainly needed in cases of segmental bone defects, nonunions, resections due to cancer and spinal fusion procedures. The commercial product of BMP2 is mostly used for spinal fusion and to a much lesser extent for long bones and maxillofacial bone repair. This could be explained by the high cost of the BMP2 commercial product and its associated side effects, mostly in spine surgery, or the notion that current therapeutic solutions, namely autografts and distraction osteogenesis, provide acceptable clinical results. In order to overcome this hurdle, there is a need for data showing that BMP gene products could be cost-effective and eliminate the side effects of currently used treatments. Moreover, BMP gene therapy could be even more attractive if proven to be a preventive treatment for the development of nonunions, a condition which cannot be predicted with current prognostic means.

Although a wide variety of means for gene delivery exist, as shown above, it is conceivable that very few would be approved for human use. Hence, there is a need to consider all the pros and cons of the different methods and proceed with the most promising approach. One of the important considerations would be the efficacy of bone regeneration. Apart from defining which BMP is the most potent bone inducer, one should consider combining the BMP gene with an addition BMP gene, an angiogenic factor like VEGF, a suppressor of BMP antagonists, or a chimera of BMP and its receptor, as was recently described [93].

Most of the studies cited in this review relate to bone regeneration. There is no doubt that cartilage regeneration is currently considered an unmet clinical need. Only a few studies have shown that BMP gene delivery could lead to articular cartilage repair. In most reports, the overexpression of various BMPs results in hypertrophic cartilage formation that leads to ossification. Hence, it is more likely that other factors or transcription factors will need to be overexpressed or downregulated to induce chondrogenesis.

Finally, with the approval of several viral gene therapies, such as LUXTURNA® for the treatment of retinal dystrophy [94] and CD19-targeted chimeric antigen receptor (CAR) T-cell immunotherapies, in which CD19 is inserted into the patient’s own T-cells in vitro [95], there is hope that we will soon enter the age of approved gene therapies for skeletal tissue regeneration.

Abbreviations:

AAV

adeno-associated virus

BMP

bone morphogenetic protein

CAR

chimeric antigen receptor

GAM

gene-activate matrix

hMDSC

human muscle-derived stem cell

miRNA

microRNA

MSC

mesenchymal stem cell

PPARγ

peroxisome proliferator-activated receptor γ

mRNA

messenger RNA; PLGA, poly(lactic-co-glycolic) acid

TGF-β

transforming growth factor-β

VEGF

vascular endothelial grown factor

3D

three-dimensional