The role of growth factors in stem cell-directed chondrogenesis: a real hope for damaged cartilage regeneration

Chondrogenesis proceeds in two main stages: condensation and differentiation. During the condensation stage, the precartilaginous mesenchyme is divided into chondrogenic and non-chondrogenic domains. The condensation starts with cell movement followed by an increase in cell-packing density. This process is associated with enhanced cell-to-cell contacts and mutual interactions due to the adhesion molecules, Ca² - dependent N-cadherin and Ca²⁺- independent N-CAM, as well as the gap junctions [13].

The differentiation stage comprises cell-matrix interactions, facilitated by the binding of fibronectin to syndecan. The increased cell proliferation and ECM remodeling is associated with the appearance of tenascin, matrilins and thrombospondins, such as, for example, cartilage oligomeric protein (COMP), as well as with the disappearance of type I collagen, fibronectin and N-cadherin. These events result in the transformation of chondroprogenitor cells into fully differentiated chondrocytes [14], while the prechondrogenic clusters develop into the ECM components, including type II collagen and aggrecan [15]. When chondrocytes enlarge, they enter the hypertrophic phase. In this phase the ECM is enriched in type X collagen and fibronectin. Then the chondrogenic cells change their environment through mineralization by calcium salts. During endochondral ossification, the chondrocytes undergo apoptosis followed by blood vessel invasion. Next, bone osteoprogenitors and the cartilagenous matrix are replaced by the mineralized matrix [16]. Another significant event is the change in fibronectin expression, as three different isoforms of this glycoprotein emerge [17]. The most suitable markers of the stages in this process are the various collagen types, with expression of type I, III and V collagens specifically marking condensation of the mesenchymal cells. Chondroprogenitor cell differentiation reveals the expression of cartilage-specific collagens (types II, IX and XI). Ultimately, the proliferating chondrocytes express type VI collagen as well as matrilin 1, while type X collagen indicates the hypertrophic zone [18].

Chondrogenesis is also connected with morphological changes of the cells. The chondroprogenitors, which are elongated and fibroblast-like, transform into spherical chondroblast- and chondrocyte-like cells [19]. The changes occurring at the cellular level result from the events taking place at a molecular level.

The principle regulators of chondrogenesis are the members of transforming growth factor β (TGF-β) superfamily, fibroblast growth factors (FGFs), insulin-like growth factors (IGFs), and members of the wingless-type (Wnt) signaling pathway [20]. The TGF-β superfamily members regulate cell proliferation, differentiation and apoptosis during embryogenesis. They include the secretory isoforms: TGF-β1, -β2 and -β3, with TGF-β3 having the highest chondrogenic potential of all isoforms, and their action results in rapid cell differentiation [21]. The TGF-βs and BMPs bind to the extracellular domains of specific receptors containing serine/threonine kinase activity in their intracellular domains, and require Sma and Mad related (SMAD) proteins for signal transduction within the cells (Fig. 1). Upon binding of the ligand to the receptor, homodimeric complexes of the receptor are formed, resulting in one subunit phosphorylating its partner subunit on serine/threonine residues. This starts a cascade of events involving phosphorylation of SMAD proteins. The TGF-β pathway requires SMAD 2 and SMAD 3, whereas the BMP signaling is dependent upon SMAD 1, 5 and 8. The phosphorylated receptor-regulated Smads (R-SMAD) reacts with SMAD4 to create a heterocomplex. The complex, consisting of R-SMAD and common-mediator Smad (Co-SMAD), enters the nucleus and by binding to promoters of target genes, controls their transcription (Table 1) [22].

Another way to regulate many cellular processes employs the mitogen-activated protein kinase (MAPK) pathway using the extracellular signal regulated kinases (ERKs), the C-Jun-NH₂-terminal kinases (JNKs) as well as the p38 MAPK (Fig. 2) [23]. Protein kinase A (PKA) potentiates these effects, acting via phosphorylation of Ser/Thr residues of crucial substrates, as well as by indirectly regulating the expression of cyclic Adenosine 3, 5′-monophospahte (cAMP) responsive genes via the cAMP response element binding protein (CREB), which binds to the CRE sites in the promoters of target genes. PKA also regulates chondrogenesis through activation of protein kinase C alpha (PKCα) [19] in cytosol and membrane fractions of mesenchymal cells during chondrogenesis [24]. In addition, PKA phosphorylates transcription factor Sox9 (sex determining region Y box 9) resulting in its transfer to the nucleus and augmentation of the transcriptional activity of these genes [22, 23]. The expression of Sox9 is up-regulated by sonic hedgehog transcription factor (SHH), which induces Bapx1 expression, which, in turn, indirectly up-regulates Sox9 expression. In cooperation with BMPs, a positive feedback loop is formed between Bapx1 and Sox9, which influences the expression of chondrogenic factors [25].

Other members of the Sox family (L-Sox5 and Sox6) are also involved in chondrogenesis. Sox9 and Runt-related transcription factor2 (Runx2) are the principal regulators, but they differ in function; while Sox9 leads to articular cartilage formation, Runx2 is involved in hypertrophic maturation of the cells [18]. Both factors are expressed throughout the entire chondrogenic process, starting from mesenchymal condensation and ending with terminal chondrocyte hypertrophy. The main function of Sox9 is to guarantee cell survival by activating the expression of type II collagen, alpha I gene (Col2a1), and other early cartilage marker genes [26]. Sox9 acts through its carboxyl terminal domain, which interacts with the CREB-binding protein (CRBP)/p300 complex, inducing the expression of specific chondrogenic genes [27].

The influence of BMP subfamily signaling polypeptides results in various biological responses such as cell growth, proliferation, differentiation, and apoptosis [28, 29]. BMP-2 is commonly found in cartilage and bone but generally not in tendons [30]. It increases binding of the NF-Y/p300 complex to Sox9 promoter, which results in Sox9 accumulation. In addition, BMP-2 is involved in histone hyperacetylation and methylation of the Sox9 promoter [31]. Recombinant human BMP-4 induces early chondrogenesis through the mediation of Runx2, while BMP-2 and BMP-7 stimulate cellular condensation. Moreover, BMP-4 stimulates the synthesis of type II collagen, aggrecan, and the ECM proteins. It is worth mentioning that BMPs, apart from promoting chondrogenesis, are able to stimulate endochondral ossification [32]. BMP-1, -2 and -3, differential splicing products of the Bmp-1 transcript, in combination with BMP-7 (known as osteogenic protein-1 [OP-1]), enhance bone formation via osteoblast production and control and remodel the ECM proteins whose distribution depends on matrix metalloproteinase (MMP) activity. However, despite the increased use of BMPs to improve bone healing, it was recently reported that administration of human BMP-7 and autologous bone grafting is no better than autologous grafting alone in patients with congenital pseudoarthrosis of the tibia [33].

The increase of Bmp-1 expression results in an increased expression of type I collagen and osteocalcin [34]. BMP-6 is a very effective regulator of MSC differentiation into osteoblasts during both development and adulthood. The endogenous expression of BMP-6 is up-regulated by the EGF-like factor and might contribute to the creation of bone-forming cells from MSCs [35]. BMP-7 is a crucial factor in skeletal development. Under pathologic conditions, BMP-7 plays a protective role for normal and OA chondrocytes by improving their survival and anabolic properties and by its ability to repair damaged cartilage [36]. The cartilage-derived morphogenetic proteins CDMP-1, -2, and -3 (also known as BMP-14, -13 and -12, respectively) are responsible for both evolution of cartilaginous tissue during early limb development and the formation of the articular joint cavity during development. Moreover, they also play a very important role in the regeneration and maintenance of the articular cartilage [37].

Mammalian fibroblast growth factors (FGFs) are a family of proteins considered paracrine regulators engaged in tissue patterning and organogenesis during embryonic development. FGF-18 has an anabolic effect on cartilaginous tissue, while FGF-2 participates in cartilage homeostasis [38]. The Wnt signaling pathway influences chondrogenesis via regulation of FGF synthesis. The type of response depends on the Wnt-2a and Wnt-2c proteins, which induce FGF-10 and FGF-8 at early stages of development. The members of the FGF family are engaged in limb initiation and limb bud outgrowth. The most important FGF family member affecting chondrogenesis is FGF-2, which functions as an effective mitogen for cell types of mesenchymal origin. Although the action of FGF-2 in chondrogenesis is not completely understood, it has been suggested that this factor might enhance the differentiation of human MSCs into chondrocytes [39]. Activated forms of FGF-10 and FGF-8 act in positive feedback loops. FGF-10 affects the expression of Wnt-3a, thus influencing β-catenin synthesis, and maintains its own expression via FGF-8 [40]. Homeobox (Hox) transcription factors are essential for activating expression of FGF-8 as well as SHH genes. The role of these factors, encoded by the HoxA and HoxD gene clusters, is crucial for the condensation stage of chondrogenesis [14].

IGFs play an important role in skeletal development. They stimulate cell proliferation, and regulate apoptosis and expression of chondrogenic markers. The cellular response to IGFs requires type I tyrosine kinase receptor (IGFIR). IGF binding results in activation of four signaling pathways: MAP kinase, extracellular signal-regulated kinase-kinase 1/2 (MEK 1/2), extracellular signal-regulated kinase 1/2 (Erk 1/2) and phosphatidylinositol-3-kinase-Akt (P13K-Akt) [41].

Osteoarthritis (OA) appears when an imbalance between the synthesis and degradation of ECM components occurs, primarily type II collagen, and aggrecan. MMPs are also involved in this process [42]. FGF-18 and FGF-9, in association with fibroblast growth factor receptor 1 (FGFR-1), induce the production of vascular endothelial growth factor (VEGF) and its receptor, VEGFR-1, thus promoting vascular invasion of the tissue in the prehypertrophic and hypertrophic zones [40]. Type II collagen is the best-known and most useful marker of the chondrocyte phenotype. Its synthesis is programmed by two RNA isoforms—the long (IIA) and the short (IIB) isoforms—generated by alternative splicing. The long form is characteristic of pre-chondrocytes, and the short one of mature chondrocytes [43]. The end of the proliferative stage and the emergence of the post-proliferative stage is accompanied by the presence of parathyroid hormone-related peptide (PTHrP), expressed in the periarticular region by proliferating chondrocytes and perichondrial cells. The PTHrP receptor is not expressed in undifferentiated cells but it is up-regulated in MSCs that undergo chondrogenic differentiation [44]. The Indian hedgehog (IHH) protein likely stimulates PTHrP production in the periarticular growth plate, thus delaying the switch from the proliferative to the hypertrophy phase. Moreover, IHH supports chondrocyte proliferation via PTHrP-independent arrangements [45]. The HH pathways are involved in chondrogenesis, chondrocyte proliferation and differentiation, as well as in OA development. IHH and SHH are ligands of the HH pathway. SHH signaling induces BMPs pathways and controls MSC differentiation towards the chondrogenic lineage in vivo [46]. Better control of the chondrogenic process in connection with the use of stem cells can contribute to more successful treatment of OA.

Evans [47] reviewed studies that used gene therapy to repair damaged bones and cartilage, both in animal models and in humans, using either non-viral or viral vectors. The author found that several major strategies were successful, as follows: (i) direct injection of DNA or vectors containing genes encoding growth factors into injury sites; (ii) the use of genetically modified allogeneic cell lines; and (iii) intra-operative tissue harvesting (either intact or isolated cells) followed by transduction with viral vectors and return to the injury site. Furthermore, Evans concludes that chondrocytes, either transduced or transfected by DNA cloned in an appropriate vector, stimulate chondrogenesis [47].

MSCs are multipotent stromal cells that can differentiate into a number of different cell types. MSCs are non-hematopoietic cells that originate from several sources, although the best-characterized populations derive from bone marrow (BM-MSCs). Besides BM-MSCs for cartilage repair cells have been isolated from various tissues such as synovium, periosteum, adipose tissue or synovial fluid [12, 48, 49]. Among these cell lines, synovium-derived MSCs (SMSCs) have been recently recognized as an appropriate cell type with chondrogenic potential and high proliferative capacity [49]. In vitro studies showed that the chondrogenesis process of SMSCs could be successfully conducted using TGF-β1 and BMP-2 [50]. Moreover, SMSCs could be also easily harvested, for example, during an arthroscopic examination of the joint. Chondrocytes derived from the MSC are suitable for application in regenerative medicine, since they are also easy to isolate and manipulate [48]. During maintenance, proliferation, migration and subsequent differentiation, MSCs are involved in the reciprocal transfer of many signals with neighbouring cells [51‐53]. Under appropriate culture conditions, MSCs can differentiate into fibroblasts or adipogenic, osteogenic, and chondrogenic lineages [54]. The classification of MSCs is troublesome, due to both a lack of a clear definition and generally accepted selective markers. Nevertheless, the basic identification standards include the following: adherence properties, expression of cell surface antigens (such as CD29, CD44, CD73, CD90, CD105, CD 106 and CD 166), and the absence of hematopoietic markers [46, 55]. Although chondrogenesis in vitro seems to be relatively simple, it was shown that during directed differentiation of MSCs up to 300 genes are misregulated [56].

Stem cell homing involves an active recruitment of endogenous cells that can be found in the tissue niche. Endogenous stem cell migration can occur in two ways. The first is through the circulatory system, aided by interactions with microvascular endothelial cells at a target site. This process includes stem cell rolling, adhesion, chemotaxis, and invasion. The second way, which is independent of blood flow, involves cell trafficking and requires amoeboid cell movement [57]. Unfortunately, MSCs do not migrate easily in the bloodstream. For this reason, translocation of MSCs from bone marrow to peripheral blood is a challenge [58]. MSCs are suitable for allogeneic transplant due to their immunoprivileged properties. In MSCs, the expression of major histocompatibility complex I proteins (MHC I) is low and there is no expression of MHC II, which decreases the risk of transplant rejection [59]. Apart from revealing multipotency, MHCs can modulate immune response and inflammation. In addition, they release trophic, anti-inflammatory, and immunomodulatory factors that exhibit pro-regenerative properties [60]. Moreover, in contrast to mature chondrocytes, MSCs can be expanded ex vivo to sufficiently high numbers, making them an abundant cell source for autologous cell therapies. Unfortunately, it is difficult to obtain a pure population of stem cells, as well as precisely control the direction of MSC differentiation [61]. Another disadvantage is the relatively low percentage of MSCs in bone marrow; moreover, in the elderly, the cell proliferation rate of isolated MSCs is often insufficient. Finally, in OA patients, isolated MSCs have a lower differentiation rate, which is another obstacle that needs to be overcome before the technique can become more widely used [62].

Induced pluripotent stem cells (iPSCs) possess the capacity to develop into derivatives of all three primary germ layers (endoderm, mesoderm, and ectoderm). They have unlimited self-renewal capacity, high developmental plasticity, and reduced immunogenic properties. These characteristics explain the enormous interest in these cells, which can be used as patient-specific cells to repair musculoskeletal tissues that have limited regeneration potential, such as bone or cartilage [63]. Most human iPSCs described in the literature originate from fibroblasts. They are created by reprogramming involving forced expression of transcription factors Sox2, Oct3/4, Klf4, and Myc in somatic cells [64]. The coordinated expression of pluripotency genes and/or an oncogene causes the re-establishment of the transcriptional information and cell morphology [65]. However, these similarities do not indicate that iPSCs are equivalent to embryonic stem cells (ESCs) at a molecular or functional level. As a result, it is necessary to further examine these cells, as well as the methods used for their reprogramming. The first studies on the induction of iPSCs were based on transduction of the cells with viral oncogenes. However, this approach increases the risk of insertional mutagenesis and tumor formation. Current methods to induce iPSCs rely upon elimination of permanent transgene integration, Cre-Lox mediated reprogramming, or mRNA/protein transfection [66]. Unfortunately, non-integrating methods suffer from extremely low transduction efficiency, which amounts to 0.001% in the case of adenoviruses [67]. ESCs will not be discussed here because they will likely never be used in clinical medicine due to ethical controversies. In contrast, the application of iPSCs is non-controversial. Moreover, iPSCS have another advantage over ESCs in that the genetic information derives from the patient’s genome and thus immune rejection is less probable [68]. However, the major drawback in the use of iPSCs for tissue engineering is the difficulty in obtaining a uniform cell population in the tissue of interest. This creates the danger of teratoma formation from undifferentiated cells [60]. Another drawback is the very low yield of the cells, together with the fact that they do not emerge in culture until three weeks after transduction [69]. It seems likely that iPSCs may revolutionize tissue engineering, and research on iPSCs is developing very quickly; however, this remains an emerging field and much more research is required. Potential advantages and disadvantages of the use of MSCs and iPSCs for transplantation are presented in Table 2.