Inhibitory function of parathyroid hormone-related protein on chondrocyte hypertrophy: the implication for articular cartilage repair

Cartilage repair tissue is usually accompanied by chondrocyte hypertrophy and osseous overgrowths, and a role for parathyroid hormone-related protein (PTHrP) in inhibiting chondrocytes from hypertrophic differentiation during the process of endochondral ossification has been demonstrated. However, application of PTHrP in cartilage repair has not been extensively considered. This review systemically summarizes for the first time the inhibitory function of PTHrP on chondrocyte hypertrophy in articular cartilage and during the process of endochondral ossification, as well as the process of mesenchymal stem cell chondrogenic differentiation. Based on the literature review, the strategy of using PTHrP for articular cartilage repair is suggested, which is instructive for clinical treatment of cartilage injuries as well as osteoarthritis.


Introduction
Articular cartilage injuries and osteoarthritis (OA) are commonly encountered in joint diseases. Current treatments aim at generating hyaline-like repair tissue with a stable, permanent chondrocyte phenotype. However, the repair tissue is often accompanied by chondrocyte hypertrophy and bony outgrowths, in particular with respect to bone marrow-eroding techniques [1,2]. Th e progressive abnormal hypertrophy may result in degradation of the matrix, impairing the function of the repair tissue. A role for parathyroid hormone-related protein (PTHrP) in regulating endochondral ossifi cation by inhibit ing chondrocytes from hypertrophy has been demon strated. Moreover, the inhibitory eff ect of PTHrP on expression of hypertrophy-related markers in articular chondrocytes and during chondrogenic diff erentiation of mesenchymal stem cells (MSCs) has been reported by a number of studies [3][4][5][6][7]. Th ose fi ndings indicate that PTHrP has potential for cartilage repair through inhibiting chondrocyte hypertrophy.
Th is review presents, compares and discusses the inhibitory function of PTHrP on chondrocyte hypertrophy in articular cartilage and during the process of endochondral ossifi cation, as well as the process of MSC chondrogenic diff erentiation. Moreover, a strategy for using PTHrP for articular cartilage repair is suggested.

Chondrocyte hypertrophy
Chondrocyte hypertrophy is commonly found during both endochondral ossifi cation and the process of articular cartilage repair. Th e former is a physiological process of bone formation, during which chondrocytes become larger and produce collagen type X ( Figure 1A,B). Th ese cells are called hypertrophic chondrocytes (quite diff erent from normal chondrocytes, which secrete and maintain the cartilaginous matrix rich in collagen type II and aggrecan). Hypertrophic chondrocytes mineralize surround ing matrix, secrete vascular endothelial growth factor to induce blood vessel formation, and fi nally under go apoptosis. Th en osteoblasts fi ll up the vacancy left by hypertrophic chondrocytes and synthesize bone matrix, a process that results in new bone formation [8]. Because chondrocytes go through a series of orderly changes during endochondral ossifi cation, the cartilage is referred to as 'transient' cartilage. In articular cartilage, hypertrophic chondrocytes usually exist quiescently in calcifi ed layers under the tidemark [9] (Figure 1C,D). Th e upper healthy articular chondrocytes maintain a stable phenotype and are resistant to hypertrophic diff erentiation [10]. So articular cartilage is called 'permanent' cartilage, which maintains the structure of functional hyaline cartilage throughout life. Under pathological conditions, however, such as cartilage injuries and OA, chondrocyte hypertrophy can be reactivated in the repair process ( Figure 2). Th e normal articular chondrocytes thus enter hypertrophic diff erentiation, resembling the process of endochondral bone formation [11,12], which can cause the upper cartilage to become calcifi ed and a relatively thin repair tissue to be formed [1,9]. Hypertrophic chondrocytes in endochondral ossifi cation and cartilage repair tissue show similar characteristics, like the production of collagen type X [13] and matrix metallo proteinase 13 (MMP-13) [14], promoting degrada tion of the cartilage matrix. Th e close link between aberrant hypertrophy and the inferior quality of cartilage repair tissue indicates that chondrocyte hypertrophy might be a potential therapeutic target to improve cartilage repair.

Biology of PTHrP
PTHrP is a protein member of the parathyroid hormone (PTH) family. It was first identified as a factor required for humoral hypercalcemia of malignancy [15]. PTHrP has been found in many tissues -not only in highabundance sites such as breast, hair follicles and cartilage, but also in previously unrecognized sites, including enthesis and periosteum [16]. PTHrP is a 141 amino acid polypeptide, and most of its biological functions are mediated by its amino terminus, such as the eff ect on cartilage [17]. Its homolog, parathyroid hormone (PTH), has the same amino-terminal 34-residue peptide fragments as PTHrP, and they thus share a common receptor, parathyroid hormone 1 receptor (PTH1R) [18]. Th e high similarity in the functional domain makes PTH and PTHrP equally potent at inhibiting chondrocyte hypertrophy, which has been proved by some studies [19][20][21]. Th us, information obtained from PTH research on hypertrophy inhibition [3,[22][23][24] is also applicable for PTHrP.

The role of PTHrP in endochondral ossifi cation
During endochondral ossifi cation, chondrocytes move through an orderly diff erentiation program: periarticular proliferating chondrocytes, columnar proliferating chondro cytes, prehypertrophic chondrocytes and hyper trophic chondrocytes. Th e role of PTHrP in endochondral ossifi cation was initially studied about 20 years ago. PTHrP is secreted by periarticular proliferating chondrocytes, while its receptor, PTH1R, is mainly expressed in prehypertrophic chondrocytes [25]. When the PTHrP gene was knocked out via homologous recombination in murine embryonic stem cells, the mice died shortly after birth and showed abnormal endochondral bone development [26]. Th e absence of PTHrP caused diminished chondrocytes and accelerated hypertrophic diff erentiation leading to premature mineralization of extracellular matrix and apoptosis [27]. Targeted overexpression of PTHrP under the control of the cartilage-specifi c collagen type II promoter resulted in the opposite eff ect of chondrodysplasia through delay of the terminal diff erentiation of chondrocytes, inhibition of apoptosis and disruption of endochondral ossifi cation [28]. Because Bcl-2, an anti-apoptotic protein, lies downstream of PTHrP [29], overexpression of PTHrP increased the expression of Bcl-2, inhibiting apoptotic cell death and disrupting growth plate architecture as well. Similar results were obtained when the PTHrP receptor was manipulated genetically [30,31].
Another important factor involved in endochondral ossifi cation is Indian hedgehog (Ihh), which is expressed by prehypertrophic chondrocytes [32] and acts in con junction with PTHrP to modulate endochondral ossifi cation. Th e foundational discovery of the Ihh-PTHrP regulatory axis was made by Vortkamp and colleagues in 1996 [33]. Th ey found that Ihh stimulated proliferating chondrocytes to produce PTHrP, which in turn acceler ated the proliferation of periarticular cells and prevented the onset of chondrocyte hypertrophy, fi nally keeping chondrocytes in a proliferating state. Th is negative feedback loop regulates the balance between proliferation and maturation of chondrocytes, ensuring orderly bone formation [33].
Th e prevailing paradigm -Ihh is mediated by PTHrP to regulate chondrocyte hypertrophy indirectly (it is also the best characterized regulatory function of PTHrP)gained wide acceptance for more than a decade until Kobayashi and colleagues [34] and Mak and colleagues [35] made a breakthrough. Using PTHrP and Ihh mutant mice, Kobayashi and colleagues found that decreasing or enhancing Ihh caused delayed or accelerated periarticular chondrocyte diff erentiation and reduced or increased the length of the columnar region, respectively, with these eff ects remaining unchanged regardless of whether PTHrP signaling was maintained or disrupted. Th erefore, they concluded that Ihh could stimulate periarticular chondrocyte diff erentiation and cause elongation of the columnar region independent of PTHrP [34]. Based on their results, Mak and colleagues further revealed a novel role of Ihh in regulating chondrocyte hypertrophy without the infl uence of PTHrP. During endochondral bone formation, PTHrP-dependent Ihh signaling inhibiting chondrocyte hypertrophy is dominant, thereby obscur ing the promoting eff ect of PTHrP-independent Ihh signaling [35] ( Figure 3A). Moreover, they speculated that canonical Wnt and bone morphogenetic protein (BMP) signaling may contribute to this non-canonical pathway as well [35].
Besides Ihh, PTHrP can be regulated by other factors as well. Amano and colleagues [36] reported that Sox9 family members inhibited the late stages of endochondral ossifi cation by up-regulating the expression of PTHrP. Human MSCs transfected with the SOX9 gene exhibited enhanced PTHrP expression together with reduction of hypertrophic markers [37]. However, the regulatory eff ect of SOX9 on expression of PTHrP diff ered depending on the target organs in vivo [38]. It has also been welldemonstrated by in vitro and in vivo studies that insulinlike growth factor 1 (IGF-1) signaling suppresses PTHrP expression, and thus modulates growth plate development [39]. Besides, the canonical Wnt pathway is known to promote chondrocyte hypertrophy via inhibition of PTHrP signaling activity instead of PTHrP expression [40].
So how does PTHrP function to inhibit premature hypertrophy of chondrocytes? Th e intracellular pathways have not been completely clarifi ed yet. Th e PTHrP receptor PTH1R is a classical G-protein-coupled receptor, which transduces signals via the G s or G q/11 family and then activates the cAMP/protein kinase A (PKA) o r phospholipase C (PLC)/protein kinase C (PKC) pathway [41]. It has been shown that the inhibitory eff ects of PTHrP preventing precocious chondrocyte hypertrophy are primarily facilitated by the activation of cAMP/PKA signaling downstream [42]. Chondrocyte hypertrophy could be delayed in mice with a PTH1R mutation that specifi cally interrupted the PLC/PKC pathway but did not aff ect cAMP/PKA activation [43]. In chimeric mice with disrupted Gnas exon 2 (encoding G sа ), chondrocytes prematurely undergo hypertrophy, resulting in a phenotype similar to that of the PTH1R -/cells [44]. Taken together, PTHrP favors the G sа /cAMP/PKA-dependent signaling pathway to inhibit hypertrophic diff erentiation. Several studies have focused on the downstream signaling pathways of PKA. Kozhemyakina and colleagues [45] found that protein phosphatase 2A (PP2A), which is activated by PKA, promoted dephosphorylation of histone deacetylase 4 (HDAC4), which was translocated into the nucleus and repressed the activity of MEF2 transcription factors, ultimately attenuating the rate of chondrocyte hypertrophy. Also, Correa and colleagues [46] showed that Zfp521, the zinc finger transcriptional coregulator, interacted with HDAC4 in the nucleus and this complex repressed expression of runx2-mediated target genes. Th ey generated mice with Zfp521 condition ally deleted in chondrocytes, resulting in early hypertrophic transition and reduced growth plate thickness. Th is phenomenon is similar to that in the PTHrP -/and PTH1R -/mice as well. Th eir recent research indicated that deletion of Zfp521 from chondrocytes rescued Jansen metaphyseal chondrodysplasia, a disorder caused by a constitutively activating mutation of PTH1R [47]. Furthermore, PTHrP inhibits runx2 expression in chondro cytes via Nkx3.2/Bapx1-mediated repression [48] or cyclin-D1-CDK4-induced phosphorylation and degradation of runx2 and runx3 [49]. All the pathways mentioned above are transduced via the cAMP/PKA pathway. Several studies have also implicated the PLC/PKC pathway in PTHrP's inhibitory function. Chen and colleagues [50] reported that p38 mitogen-activated protein kinase (MAPK) induced the expression of Bcl-2 as well as collagen type X, and PTH1R activation could block p38 MAPK activity, which was mediated by PKC instead of PKA. Th e above-mentioned pathways are summarized in Figure 3B. Further investigations are needed to improve our understanding of the mechanisms by which PTHrP inhibits chondrocyte hypertrophy.

The role of PTHrP in articular cartilage maintenance
Based on the well established theory of PTHrP's role in endochondral ossifi cation, researchers have also tried to apply it to articular cartilage to better understand its pathogenetic mechanisms. A few publications demonstrated that PTHrP could regulate articular chondrocytes. In 2008, Chen and colleagues found that PTHrP participated in the maintenance of articular cartilage as it did in growth plate cartilage [51], using a PTHrP-LacZ knockin mouse that can be used to characterize PTHrP gene expression [18]. Th e PTHrP-producing cells, in both growth plate cartilage and articular cartilage, are derived from the same source, the chondroepiphysis, which is divided by the secondary ossifi cation center into two distinct PTHrP-expressing subpopulations -the chondro cytes of the articular surface and periarticular proliferating chondrocytes of the growth plate [16] -while PTH1R is expressed in the underlying prehyper trophic chondrocytes.
Jiang and colleagues [4] used a co-culture model to evaluate the interaction of chondrocytes derived from diff erent layers of articular cartilage. Th ey demonstrated that PTHrP produced by the superfi cial or middle layer could block the alkaline phosphatase activity and related mineralization of chondrocytes from the deep layer. Th is study provides in vitro evidence that, in healthy articular cartilage, PTHrP secreted by chondrocytes from surface layers inhibits the hypertrophic potential of chondrocytes residing in the deep layer so as to maintain the homeostasis of articular cartilage, but the eff ect was not confi rmed in vivo. More convincing evidence was provided by conditional deletion of PTHrP in midregion articular chondrocytes via growth and diff erentiation factor 5 (Gdf5) control sequences. After destabilizing medial meniscus, the knock-out mice exhibited severe cartilage degeneration, indicating that PTHrP plays an important role in the physiological regulation of articular cartilage maintenance [52].
Since PTHrP shows inhibitory eff ects on articular chondrocyte hypertrophy physiologically, the role of PTHrP under pathological conditions was explored as well. It was reported that the expression of PTHrP in articular cartilage was 5.42-fold higher than that of osteophytic chondrocytes, which represents a prototype of cartilage repair tissue [53]. It is probable that the decreased expression of PTHrP is insuffi cient to inhibit abnormal hypertrophy in repair tissue; therefore, supplement of exogenous PTHrP to diseased cartilage is of great signifi cance. Wang and colleagues [54] transfected bovine articular chondrocytes with human PTHrP (hPTHrP) constructs and applied cyclic tensile strain to induce arthriti c changes simultaneously. Overexpression of hPTHrP inhibited cyclic tensile strain-induced collagen type X expression, suggesting the involvement of PTHrP in resisting mechanical strain-induced hyper trophic-like changes. In another study, human articular chondrocytes were treated with azacytidine to induce terminal diff erentiation, mimicking the situation in OA [3]. Treatment with exogenous PTH (1-34) (an aminoterminal fragment of PTH) signifi cantly elimi nated the increased expression of collagen type X, alkaline phosphatase and Ihh induced by azacy tidine. More over, when papain-induced OA rats were injected intraarticularly with PTH (1-34) for 5 weeks, the OA cartilage was almost restored to a normal state. It should be noted that, in this research, PTH did not exert any adverse eff ects on normal chondrocytes or healthy joints, only on OA-aff ected cartilage. However, studies from Kudo and colleagues [55] showed that bone marrow-derived MSCs transformed from a chondrogenic to a fi broblastic phenotype and osteochondral defects were never covered with cartilage after 4-week PTH (1-84) treatment. Th is contradiction probably results from the diff erences in the animal models, injury types and PTH fragments in these two studies. Our group found that the time window for PTHrP administration is of great importance, which could infl uence chondro genesis and chondrocyte hyper trophy of repair tissue. More details are given in the next section.
Th e canonical Ihh-PTHrP pathway has been commonly invoked when discussing results in most publications investigating the mechanisms by which PTHrP regulates articular cartilage [3,4]. Chen and colleagues, however, raised a diff erent and challenging possibility [51,52]: in articular cartilage, the master regulatory factor is mechanical loading, rather than Ihh. Such loading would induce expression of PTHrP in the articular chondrocytes from the superfi cial and middle layers, and then transduction of the PTHrP to inhibit terminal diff erentiation and promote prolifera tion, a downstream signaling pathway similar to that of endochondral bone formation. It is surprising that the Ihh-PTHrP regulatory axis is uncoupled in articular chondro cytes. In injured and OA cartilage, tensile stiff ness in the superfi cial layer of articular cartilage decreased sub stantially [56]. It is very likely that the alteration of mechanical loading infl uences the primary responder, PTHrP, and fi nally impacts the downstream regulatory system. Although this system has not been fully confi rmed, it provides a new idea to study the mecha nisms of PTHrP and the pathogenesis of chondrocyte hypertrophy in cartilage repair.

The role of PTHrP in MSC chondrogenic diff erentiation
Large and unconfi ned cartilage injuries are nearly impossible for the eroded articular cartilage to repair by itself. Autologous chondrocyte implantation may be a practical approach to solve this problem and has been applied to the clinical treatment of cartilage defects. However, during the in vitro expansion period for collection of suffi cient cells, chondrocytes often undergo rapid dediff erentiation, which limits the clinical application of autologous chondrocyte implantation [57]. MSCs are considered an alternative cell source for cellbased therapies of cartilage injuries. Th ey can be isolated from diff erent tissues, expand rapidly and stably, and diff erentiate into chondrocytes eff ectively. It has been reported, however, that induction of chondrogenesis of MSCs in vitro is generally accompanied by unwanted hypertrophic diff erentiation [58]. Furthermore, ectopic transplantation of MSC pellets in nude mice was followed by calcifi cation and vascular invasion, leading to cartilage phenotypic instability [59]. Because PTHrP exerts inhibitory eff ects on terminal diff erentiation of articular cartilage, some researchers proposed that PTHrP may be a candi date to inhibit hypertrophy during MSC chondrogenic dff erentiation as well.
It was reported that co-culture of human MSCs with human articular chondrocytes under chondrogenic induc tion could promote chondrogenesis and inhibit hyper trophy of the engineered cartilage [60]. Similarly, Fischer and colleagues [5] described that when MSC pellets were induced to diff erentiate in chondrocyteconditioned medium, hypertrophy was signifi cantly inhibited both in vitro and in the following ectopic transplantation study. Furthermore, when MSCs were cocultured with chon dro cytes directly, in vivo calcifi cation was completely inhibited. Further investigations were carried out to identify the chondrocyte-derived soluble factors involved in these eff ects and PTHrP was found to be the regulator; however, in situ repair studies are needed to confi rm this conclusion. In the studies using PTHrP as a supplement of the chondrogenic medium, MSC pellets diff erentiated with suppressed hypertrophy in vitro [5][6][7]. Nevertheless, the eff ect of PTHrP on chondrogenesis was discrepant. Weiss and colleagues [6] reported an inhibitory eff ect on collagen type II expression with 1 or 10 ng/ml PTHrP after 21 days of chondrogenesis, while Kim and colleagues [7] showed collagen type II and SOX9 gene expression increased up to 4-fold in bone marrow-derived MSCs and adipose tissue-derived MSCs when cells were treated with 10 or 100 ng/ml PTHrP from the 14th day of culture. Th is contradiction is probably due to the diverse treat ment time or the diff erent fragments and concen trations of PTHrP they used. Th e above phenomenon was reported in normal cells, whereas Kafi enah and colleagues [611] engineered cartilage with MSCs from OA patients. Hypertrophy was remarkably inhibited with PTHrP treatment, though the OA cells were more inclined to terminal diff erentiation. Moreover, the process of chondro genesis was not aff ected by PTHrP. Overall, PTHrP serves as a predominant factor in MSC-based approaches to promote cartilage repair; however, further work is still needed to minimize the undesirable eff ect on chondrogenesis before clinical use. According to what Weiss and colleagues [6] found, adding 0.1 ng/ml of PTHrP from day 21 could suppress collagen type X deposition without any negative eff ects on chondrogenic diff erentiation, while higher doses (10 or 100 ng/ml) or earlier treatment (from day 0) would lead to the suppression of chondrogenesis, valuable information when considering clinical application.

Application of PTHrP for articular cartilage repair
As described above, it is likely that PTHrP can be used to restrain abnormal hypertrophy in the cartilage repair process. Treatment with PTHrP would consist of two main methods: administration of recombinant protein or gene therapy using genetic manipulation (Figure 4).

Recombinant protein treatment
Recombinant PTHrP protein used in most studies includes only its functional domains, such as PTHrP (1-34) [3,5,6], to which the PTHrP receptor binds. Experi ments have been conducted where MSC pellets have been pre-treated with recombinant human PTHrP (1-34) and then the pellets transplanted subcutaneously into SCID mice [6]. Hypertrophy was inhibited during in vitro chondrogenesis but subsequent in vivo calcification was not repressed. Th is ineffi cacy is possibly due to the absence of PTHrP in ectopic transplantation sites, which implies the importance of sustained and suffi cient PTHrP for complete inhibition of hypertrophy.
Injection of PTHrP in sites can maintain the working concentration to some extent. Direct delivery of recombinant PTHrP without scaff olds, however, requires frequent injections to compensate for its rapid degradation and removal. Chang and colleagues [3] reported that injection of 10 nM PTH (1-34) intra-articularly every 3 days delayed progression of OA in rats. Kudo and colleagues [23] and Mizuta and colleagues [24] demonstrated that constant PTH administration into the joint cavity by an osmotic pump placed subcutaneously avoided the complexities of frequent injection.
Tissue engineering approaches are potent methods for cartilage repair. Biodegradable injectable scaff olds are utilized as a delivery system for the controlled release of PTHrP. Various kinds of materials have been developed to improve local delivery in cartilage tissue engineering, such as collagen-alginate [62]. Release rate and degradation time change with the scaff old materials and structures and an appropriate release system should be chosen according to the release profi le of PTHrP, the specifi city of the target region, and the actual target eff ect.
On the other hand, the treatment time must be optimized to avoid the side eff ects of early or prolonged PTHrP treatment. Mizuta and colleagues [24] demonstrated that a 2-week treatment with PTHrP for fullthickness articular cartilage defects resulted in successful regeneration, while a 4-week treatment resulted in an inferior repair, although the eff ect on hypertrophy inhibition in repair tissue was not assessed in this study. As mentioned above, our unpublished data indicate that it is quite important to choose a suitable time window for PTHrP administration post-injury. We injected 50 μg/ml hPTHrP (1-40) into rabbit joint cavity once a week at three diff erent time windows: 4 to 6 weeks, 7 to 9 weeks and 10 to 12 weeks after osteochondral defect construction. All rabbits in the three groups were sacrifi ced 16 weeks post-injury. Th e results showed that joint cartilage injected at 4 to 6 weeks after osteochondral defect construction exhibited a better morphological and histological appearance than the other groups. Th e repair tissue with PTHrP injection at 4 to 6 weeks displayed less expression of collagen type X and other hypertrophic markers than the groups without PTHrP injection or treated over diff erent time windows (unpublished data). Th ese results suggest that 4 to 6 weeks post-injury is a suitable time window to administer PTHrP for osteochondral defect repair in rabbit, but further tests are needed to determine whether this is the case for other animal models or PTHrP concentrations. Kudo and [23] colleagues administrated hPTH (1-84) continuously or intermittently for 2 weeks after creation of full-thickness defects. Th e repair tissue was found to be fi brous or fi brocartilaginous with continuous treatment after 8 weeks post-injury, while intermittent treatment resulted in restoration of hyaline-like cartilage. Based on the above-mentioned studies, we conclude that short-term and intermittent treatment using PTHrP in the early phase (but not the initiation) after injury may lead to superior cartilage repair. However, those pilot studies were done by diff erent groups under diff erent situations, and systematic research must be carried out to determine the conditions leading to the best therapeutic eff ect.

Gene therapy
Gene therapy represents another new approach, providing persistent synthesis of required proteins at target sites in vivo. Wang and colleagues [54] generated bovine articular chondrocytes that were transfected with human PTHrP constructs, demonstrating that PTHrP-transfected chondrocytes resisted mechanical strain-induced hypertrophy. Th is eff ect was not confi rmed by in vivo experiments. Th ough few studies on PTHrP gene therapy have been conducted, studies focusing on cartilage repair with other genes are numerous, making PTHrP gene transfer feasible and applicable.
PTHrP gene therapy could be carried out in two ways. Th e PTHrP gene (delivered by adenovirus, lentivirus, adeno-associated virus or other vectors) could be directly introduced to the joint space to aff ect surrounding cells, such as synovial lining cells, articular chondrocytes and MSCs [63], so that they secrete PTHrP. Alternatively, the PTHrP gene could be transfected into cultured cells, such as chondrocytes, MSCs and other cells, following transplantation of those cells to the target locations [64] for sustained production of PTHrP. In both methods, gene constructs or gene-transfected cells can be pre-mixed with tissue engineered scaff olds before transplantation to guarantee more durable delivery.
However, the problem of unrestrained PTHrP production, similar to excessive administration of recombinant PTHrP mentioned above, could occur with gene therapy, so the level of PTHrP would need to be precisely controlled. According to previous research, 1 ng/ml PTHrP was suitable for MSC chondrogenesis while higher doses suppressed collagen type II expression [6]. Th us, PTHrP expression should be confi ned to a certain level for better therapy.
Th e time window for PTHrP expression also requires consideration. According to our results from protein treatment, PTHrP administration at 4 to 6 weeks post-injury is the optimum time window. Th erefore, it would be better to switch the PTHrP gene on during this period. Various kinds of regulatory strategies have been explored to achieve temporal control of transgene expression. Besides the conventional exogenous stimuli, such as tetra cycline [65], that have been utilized for years, new types of stimuli, such as light [66] and ultrasound [67], have emerged and can be applied to control of PTHrP expression.

Conclusions and perspectives
Chondrocyte hypertrophy is found in the process of cartilage repair and endochondral bone formation.
PTHrP has long been recognized as a potent inhibitor of hypertrophic diff erentiation during endochondral ossification. Pilot studies demonstrated PTHrP's capacity to suppress abnormal hypertrophy in both the articular cartilage and MSC chondrogenic diff erentiation processes. So, it could be reasoned that PTHrP may potentially inhibit chondrocyte hypertrophy in cartilage repair tissue as well. Like other bioactive factors, therapy with PTHrP could be achieved using recombinant protein or gene manipulation. However, to develop an effi cient and applicable way for functional cartilage repair with PTHrP, several key issues need to be further investigated.
First, a suitable PTHrP concentration should be determined. Th is should be suffi cient to inhibit hypertrophy but have no adverse eff ects on chondrogenesis. Th e concentration changes depending on the animal model, PTHrP fragments used, administration time, and so on. Second, the time over which PTHrP functions best should be determined. Th is should be early, short and inter mittent to avoid side eff ects of earlier or prolonged PTHrP treatment. Th e release frequency and duration need to be optimized according to practical require ments.
Th ird, the PTHrP delivery system should be optimized, including the delivery material, methods of recombinant protein injection, as well as methods of PTHrP gene transfer.
Finally, more in vivo studies are necessary to gain more direct evidence on the effi cacy of PTHrP on articular cartilage repair and to elucidate the underlying mechanisms more clearly in a variety of animal models.