Enhanced articular cartilage regeneration with SIRT1-activated MSCs using gelatin-based hydrogel

To investigate the functional effects of resveratrol (RSV) on mesenchymal stem cells (MSCs), we treated MSCs with RSV continuously during ex vivo expansion. MSCs were continuously treated with RSV from passage (P) 0 to P5. A proliferative capacity of RSV-treated MSCs was higher than that of non-treated MSCs and similar with P1-MSCs. Continuous treatment of RSV on MSCs increased the stemness and inhibited the senescence. During chondrogenic differentiation in vitro, RSV-treated MSCs had higher differentiation potential and reduced hypertrophic maturation, which are limitations for hyaline cartilage formation. The histological analysis of micromass demonstrated increased chondrogenic differentiation potential. We further explored the therapeutic effectiveness of this method in a rabbit osteochondral defect model. A rabbit osteochondral defect model was established to investigate the hyaline cartilage regeneration potential of RSV-treated MSCs. Moreover, the cartilage regeneration potential of RSV-treated MSCs was greater than that of untreated MSCs. The expression levels of chondrogenic markers increased and those of hypertrophic markers decreased in RSV-treated MSCs compared with untreated MSCs. Sustained treatment of RSV on MSCs during ex vivo expansion resulted in the maintenance of stemness and enhanced chondrogenic differentiation potential. Consequentially, highly efficient MSCs promoted superior hyaline cartilage regeneration in vivo. This novel treatment method provides a basis for cell-based tissue engineering.


Introduction
Osteoarthritis (OA) involves cartilage damage, dysfunctional chondrocyte proliferation, and hypertrophic maturation [1][2][3][4] . Several types of therapies are currently used for cartilage regeneration, including bone marrowstimulating techniques, mosaicplasty, and cell-based therapies 5 . Cell-based therapies are increasingly used as a prospective treatment. Autologous chondrocyte implantation is commonly used for cartilage regeneration, requiring the in vitro expansion of autologous chondrocytes 6 . Nevertheless, these techniques have several shortcomings, such as their complexity, cost, and the loss of cartilage capacity 7,8 . Therefore, cell-based therapeutic approaches using mesenchymal stem cells (MSCs) have emerged for cartilage regeneration [9][10][11] .
MSCs have various advantages for clinical applications, such as anti-inflammatory and immunosuppressive effects, high effectiveness, and a lack of severe side effects 12 . Additionally, MSCs possess a high stemness capacity and thereby have the potential for multipotency, including chondrogenic differentiation. For these reasons, MSCs are a promising cell source for cartilage regeneration [13][14][15] . However, several factors limit the clinical application of MSCs. For example, hundreds of millions of MSCs are required. MSCs can be isolated from various organs, but the number of isolated cells is insufficient. To obtain sufficient MSCs, long-term in vitro expansion is necessary 16 ; however, acquiring an effective quantity of MSCs with sustained self-renewal and multi-lineage differentiation potential is difficult 17,18 . Moreover, MSCs tend to become fibro-like tissues after the induction of chondrogenic differentiation 19,20 , and they exhibit fibrolike tissue regeneration in vivo. Therefore, the identification of new environments for the development of highly efficient MSCs with an enhanced self-renewal capacity and chondrogenic differentiation potential is necessary.
Mohyeldin et al. 21 demonstrated that MSCs require physiological oxygen levels of between 2 and 8%. Under hypoxic conditions, the expression levels of hypoxiainducible factors increase, leading to the upregulation of Oct-4, Sox2, and Nanog, which are critical transcription factor for stemness 22 . Previous studies have suggested that under hypoxic conditions, MSCs have enhanced proliferative potential and stemness during in vitro cultivation 23 . In fact, antioxidants, such as resveratrol 24 , palm vitamin E 25 , butylated hydroxyanisole 26 , and butylated hydroxytoluene 27 , are known to not only suppress reactive oxygen species, but also delay disease progression and influence lipid peroxidation 28 . Butylated hydroxyanisole and butylated hydroxytoluene are frequently used as additives in the food industry 26,29 ; however, their use is restricted owing to their potential toxicity and carcinogenic effects 28 . Thus, the discovery of new, safe antioxidants that can overcome the disadvantages of chemical antioxidants is necessary.
The antioxidant resveratrol (RSV; 3,5,4ʹ-hydroxystilbene), a natural compound, is a phytoalexin produced by plants in response to environmental stress [30][31][32] . RSV has critical roles in the inhibition of oxidative damage 33 as well as in cell survival and proliferation [34][35][36] . Furthermore, RSV is known to enhance activity of sirtuin 1 (SIRT1), an NAD+-dependent lysine deacetylase. During in vitro cultivation, MSCs occasionally become senescent, leading to a loss of stemness. Previously, Yuan et al. 35 demonstrated that the knockdown of SIRT1 in early-passage MSCs using shRNA results in a loss of proliferative capacity and promotes cellular senescence. In contrast, the overexpression of SIRT1 in MSCs delays cellular senescence and cells maintain multipotency during long-term in vitro cultivation 35 . Moreover, RSV enhances the osteogenic and adipogenic differentiation potential of MSCs, but prolonged treatment of MSCs with RSV induces senescence 37,38 . Other studies have demonstrated that RSV has negative effects on MSCs during adipogenic and osteogenic differentiation 39,40 . These incompatible results regarding the effects of RSV on senescence and differentiation in MSCs may be explained by dose-or duration-dependent effects 34 . Hence, a novel RSV treatment method for MSCs is necessary. Additionally, the effect of RSV on chondrogenic differentiation has not been elucidated. In this study, we developed a novel treatment method using RSV that yields consistent results. Moreover, we evaluated the stemness and chondrogenic differentiation potential of MSCs after treatment with RSV. This novel treatment method provides a basis for cell-based tissue engineering.

Increased stemness and decreased senescence of RSVtreated MSCs
During in vitro cultivation, we continuously treated MSCs with RSV from P0 to P5 (P5-RMSC), and other cells were cultured up to P5 without RSV treatment (P5-MSC) (Fig. 1a). To investigate whether RSV treatment increases the self-renewal capacity of MSCs, we evaluated the cell morphology and proliferative potential. P5-RMSCs had a similar morphology to that of P1-MSCs, i.e., cells were small and spindle-shaped. However, P5-MSCs had a large, flat appearance (Fig. 1b). P5-MSCs had a decreased proliferative capacity over time, whereas the P5-RMSCs had a similar proliferative capacity to that of P1-MSCs (Fig. 1c). To determine the effects of RSV on senescence and stemness, we assessed protein expression levels of senescence and stemness markers such as P16, P21, P53 and Nanog, Oct4, Sox2, respectively. In P5-RMSCs, the stemness markers were up-regulated and senescence markers were down-regulated, similar to the expression levels in P1-MSCs, however, the opposite results were obtained for P5-MSCs, which exhibited the upregulation of senescence markers and down-regulation of stemness markers (Fig. 1d, e). Therefore, continuous treatment of RSV on MSCs was able to maintain the stemness and inhibit senescence, suggesting that it is a useful method for the establishment of highly efficient MSCs.

Enhancement of the chondrogenic differentiation potential in RSV-treated MSCs
To evaluate the chondrogenic differentiation potential, we performed micromass culture using P1-MSC, P5-MSC, and P5-RMSC. The mRNA levels of Sox-5, -6, -9, Col2a1, and aggrecan, which are chondrogenic markers, were higher in P5-RMSC than in P5-MSC and were similar to the levels in P1-MSC (Fig. 2a). The protein expression levels of chondrogenic markers were comparable in P1-MSC and P5-RMSC and a quantitative analysis supported this finding (Fig. 2b, c). To further demonstrate the increased chondrogenic differentiation potential of P5-RMSC, we performed SafO and AB staining. The P5-MSC group had a much smaller micromass size than those of the P1-MSC and P5-RMSC groups, which both had higher contents of glycosaminoglycan (GAG) and proteoglycan (PG) (Fig. 3a). We evaluated the expression levels of col2a1 and aggrecan in micromass pellets by immunocytochemistry. P5-RMSC had high expression levels of col2a1 and aggrecan, similar to the expression levels in P1-MSC (Fig. 3b, c).
Accordingly, P5-RMSC, which had enhanced stemness, had superior chondrogenic differentiation potential compared to that of P5-MSC. c Proliferation, as determined on day 1, 2, 4, and 6. Expression levels of senescence markers, P16, P21, and P53 (d) and stemness markers, Nanog, Oct4 and Sox2 (e) in P1-MSC, P1-RMSC, P5-MSC, and P5-RMSC, determined by western blotting. Quantitative analysis of protein expression levels of senescence (f) and stemenss (g) markers was performed. Each dot represents a mean value of triplicate estimates with one donor and horizontal lines display the means and 95% CIs of the values from different donors per group

Inhibition of the hypertrophic maturation of RSV-treated MSCs
During chondrogenic differentiation, hypertrophic maturation inhibits hyaline cartilage regeneration 19,41 . To determine whether P5-RMSC exhibited inhibited hypertrophic maturation during chondrogenic differentiation in vitro, we evaluated the both mRNA and protein expression levels of hypertrophic markers. P5-RMSC had lower expression levels of hypertrophic markers than those in P5-MSC (Fig. 4a, b). Additionally, an immunocytochemical analysis demonstrated that the expression of col10a1, a key marker of hypertrophic maturation, was significantly decreased in P5-RMSC (Fig. 4c). Consequentially, sustained treatment of MSCs with RSV enhanced the chondrogenic differentiation potential and prevented hypertrophic maturation during in vitro chondrogenic differentiation via improved stemness.

Increased cartilage regeneration potential in vivo
To determine whether continuous treatment of RSV on MSCs could enhance the cartilage regeneration capacity in vivo, we established a rabbit osteochondral defect model ( Fig. 5a). At 8 weeks post operation, we observed the gross morphology of regenerated cartilage. In the Hy/RMSC group, the surface of the defect site was almost fully filled with cartilage-like tissue, along with nearby cartilage. The Hy/RMSC group showed transparent cartilage-like tissues, whereas the defect, Hy, and Hy/MSC groups had incomplete cartilage tissue formation (Fig. 5b). These results indicate that the regeneration of intact cartilage surface tissues is possible via continuous treatment of RSV on MSCs.
Histological analysis of regenerated cartilage in the rabbit osteochondral defect model We observed regenerated cartilage at defect sites by MT and H&E staining. The MT stain showed higher collagen deposition on cartilage surface of the Hy/RMSC group compared to the Hy/MSC group (Fig. 6a). In the H&E stain, the Hy/RMSC group had chondrocyte-like cells with smooth cartilage-like tissues, which were similar in morphology to normal cartilage. However, the defect, Hy, and Hy/MSC groups showed fibrous tissue formation (Fig. 6b). Safranin O/Fast Green staining demonstrated increased GAG formation in the Hy/RMSC group; the Hy/MSC group showed slightly greater synthesis of GAG than those of the other groups (Fig. 6c). Additionally, O'Driscoll histological scoring was performed, and the Hy/MSC group had a slightly higher score than those of the defect or Hy group, and the Hy/RMSC group had a significantly higher score (Fig. 6d). These results demonstrate that the Hy/MSC group had slight differences in

Effects of cartilage regeneration for RSV-treated MSCs based on both chondrogenic and hypertrophic markers
Next, we evaluated the expression of extracellular matrix (ECM) proteins, type II collagen and aggrecan, in regenerated cartilage tissues by immunohistochemistry. The Hy/RMSC group showed higher expression levels of ECM proteins than those of the Hy/MSC group, which exhibited slight expression of these proteins. A quantitative analysis demonstrated that there was significantly higher expression of both type II collagen and aggrecan in the Hy/RMSC group than in other groups (Fig. 7a, b). Moreover, the expression of a hypertrophic marker, type X collagen, was assessed. The Hy/RMSC group showed very low expression of type X collagen, while other groups had high expression levels. A quantitative analysis showed that type X collagen expression was significantly lower in the Hy/RMSC group than in other groups (Fig. 7c). As a result, the continuous treatment of MSCs with RSV resulted in enhanced cartilage regeneration potential and the inhibition of hypertrophic maturation, which limits the use of MSCs for cell therapy 19,41 .

Discussion
In the present study, MSCs were continuously treated with RSV from P0 to P5, and the maintenance of their stemness was confirmed. Continuous treatment maintained the activity of sirtuin 1 (SIRT1). P5-RMSC had increased protein levels of stemness markers and decreased levels of senescence markers compared with those in P5-MSC. As shown in our previous study, the maintenance of SIRT1 activity sustains the expression of SOX2 37 ; thus, the stemness of MSCs could be maintained over time.
P5-RMSC had similar stemness to that of P1-MSC and improved chondrogenic differentiation potential (Figs. 2  and 3). Additionally, hypertrophic maturation was inhibited in P5-RMSC, resulting in the increased expression of type X collagen (COL10A1), matrix metalloproteinase 13 (MMP13), and alkaline phosphatase (ALP) (Fig. 4) 19,41 . To confirm the efficiency of the cartilage regenerative potential of P5-RMSC, which had similar stemness to that of P1-MSC, we established an osteochondral defect model in rabbits with a depth of 3 mm and a diameter of 6 mm, a critical size threshold for rabbit models 42 . This critical size of osteochondral defect sites results in an inability to selfheal. At 8 weeks post operation, P5-RMSC showed smoother and more intact cartilage that was well attached to nearby cartilage compared to other groups (Figs. 5 and 6). P5-RMSC had higher expression of ECM proteins and lower expression of hypertrophic maturation markers (Fig. 7). Therefore, P5-RMSC had increased cartilage regeneration potential with hyaline-like tissue formation in the rabbit osteochondral defect model when compared with P5-MSC. These results demonstrate that our innovative treatment method with RSV promotes the maintenance of stemness in MSCs over time and thus can increase the cartilage regeneration potential (Fig. 8).
During long-term ex vivo expansion, MSCs do not grow infinitely, leading to senescence 43 . Bonab et al. 17 demonstrated that bone marrow-derived MSCs exhibit  Several studies have demonstrated that long-term exposure to RSV could increase the senescence of MSCs 34 ; during osteogenic differentiation, RSV adversely affects adipogenic differentiation 45 . For the chondrogenic differentiation of MSCs, previous studies have demonstrated that treatment with RSV during the chondrogenic differentiation period enhances chondrogenic potential and inhibits inflammatory and degradative processes in cartilage via the activation of SIRT1 46 . Additionally, the treatment of chondrocytes with RSV up-regulates SIRT1 expression, but leads to hypertrophic maturation 47 . Unlike previous studies, our results showed that stemness can be maintained in P5-MSC via continuous treatment with RSV, thus maintaining their chondrogenic differentiation potential. We also confirmed the inhibition of hypertrophic maturation both in vitro and in vivo. For clinical applications, the recommended number of passages for MSCs is 3-5 48,49 . In general, MSCs at passages 1-2 have high multipotency and MSCs at passages 4-5 have low multipotency and a loss of replicative potential 50,51 . Accordingly, in our study, we continuously treated MSCs with RSV from P0 to P5 to verify the sustained stemness of MSCs at passage 5.
In our in vivo study, the defect group showed slight regeneration of cartilage-like tissues (Fig. 6). Although the osteochondral defect sizes were fixed at 3 mm in depth, during the surgical procedure, the bone marrow concentrates affected the regeneration of cartilage-like tissues in the defect group. In fact, bone marrow concentrates containing a variety of heterogenous cells and a lack of MSCs are known to contribute to fibrocartilage formation, rather than hyaline cartilage 52 . Thus, the bone marrow concentrates did not affect the cartilage regenerative potential in the Hy/MSC and Hy/RMSC groups. The Hy/RMSC group had hyaline-like cartilage tissue formation (Figs. 6 and 7).
Taken together, the continuous treatment of RSV on MSCs during in vitro cultivation resulted in sustained stemness and chondrogenic differentiation potential, similar to that of P1-MSC, thereby improving hyaline cartilage regeneration (Fig. 8). Therefore, the novel treatment method with RSV could provide a foundation for studies in cell-based tissue engineering with minimal side effects.

Isolation and identification of MSCs from human bone marrow aspirates
Bone marrow aspirates were acquired from the posterior iliac crest of 10 healthy donors after obtaining approval from the Institutional Review Board of Yonsei University College of Medicine. MSCs were isolated and cultivated as previously described 53 .

Chemical treatment of MSCs
Resveratrol (RSV; Sigma, St. Louis, MO, USA) was dissolved in ethanol (EtOH) at 1 µM. RSV was continuously applied to MSCs from P0 to P5 (P5-RMSC) and cells were subcultured as previously described 54 . MSCs were evaluated up to P5 based on previous studies indicating that MSCs at passages 3-5 are optimal for clinical applications 48,49 .

In vitro chondrogenic differentiation of MSCs
For the chondrogenic differentiation of MSCs, the micromass culture method was used, as previously reported 55 . Briefly, 10 µl of resuspended cells at a density of 1 × 10 5 cells per well was added to the center of the wells in 24-well plates. Cells were allowed to attach for 2 h and chondrogenic medium was then added. The chondrogenic medium was replenished every 3 days.

Quantitative real-time PCR (qRT-PCR)
Extraction of total RNA harvested from cells was performed using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocols. To synthesize cDNA, the Omniscript Reverse Transcription Kit (Qiagen, Hilden, Germany) was used for the reverse transcription of 2 μg of RNA. To conduct real-time PCR, 2×qPCRBIO SyGreen Mix (PCR Biosystems, London, UK) was used with 80 ng of cDNA. To quantify the relative expression levels of target genes, beta-actin (β-actin) was used as an internal control. Validated primers targeting Sox9 (P232240), IHH (P101104), and ALP (P324388) were used. Other primers were designed, as shown in Table 1, and all primers were purchased from Bioneer (Daejeon, Korea). Real-time PCR was performed using the ABI7900 (Applied Biosystems, Carlsbad, CA, USA) in accordance with a previous report 55 .

Western blot analysis
Cell pellets were lysed and quantified as previously reported 56

Animal experiments
Twenty New Zealand white rabbits (male, 3.5-4 kg; Doo Yeol Biotech, Seocho-gu, Seoul, Korea) were used to establish the osteochondral defect model according to previously methods 59 . Briefly, full-thickness cylindrical osteochondral defect sites with a diameter of 6 mm (3-mm depth) were formed on the surface of the femoral patellar groove using the osteochondral autotransfer system (OATS; Arthrex, Naples, FL, USA). For each osteochondral defect site, the following treatments were applied: None (defect), hydrogel only (Hy), Hy + P5-MSC (Hy/MSC), and Hy + P5-RMSC (Hy/RMSC). A total of 2 × 10 6 in vitro cultured MSCs were mixed with the GH polymer (5 wt%) that containing 0.0015 mg/mL of horseradish peroxidase (HRP). The hydrogel formation was occurred when the solution mixed with same volume of GH polymer (5 wt%) that containing 0.005 wt% hydrogen peroxide (H 2 O 2 ) 60 . At 8 weeks after the operation, the rabbits were euthanized and defect sites were extracted for the histological analysis. All animal experiments were approved by the Committee on the Ethics of Animal Experiments of Yonsei University College of Medicine (Permit No. 2016-0200).

Histological evaluation
The samples obtained from osteochondral knee defects and micromass pellets were fixed for 7 days and 24 h in 10% formalin, respectively. After fixation, the specimens were embedded in paraffin and then paraffin sections were deparaffinized, rehydrated, and washed with PBS. The sections were stained with Masson's trichrome (MT), hematoxylin-eosin (H&E), Safranin O (SafO), and Alcian blue (AB), as previously described 59 . The stained sections were observed using a VS 120 virtual microscope (Olympus, Tokyo, Japan). A quantitative analysis was performed using ImageJ v.1.48 (Aspire Software International, Leesburg, VA, USA). Moreover, the O'Drsicoll histological scoring system was used for the histological assessment 61 , and the regenerated cartilage was assessed by three blinded experts.

Immunocytochemistry
After washing paraffin sections with PBS, they were incubated in hydrogen peroxidase for 10 min to minimize nonspecific background staining. To detect chondrogenic markers, the sections were incubated with anti-COL2A1, anti-AGGRECAN, and anti-COL10A1 (1:100, Santa Cruz Biotechnology) antibodies at 4°C for at least 12 h. The attachment of secondary antibodies and fluorescent protein-conjugated secondary antibodies was performed following previously described methods 59 . Nuclei were stained with 4ʹ,6-diamidino-2-phenylindole (Sigma). An inverted fluorescence microscope (IX-71; Olympus, Tokyo, Japan) was used to acquire images and expression levels were quantified using ImageJ.

Statistical analysis
Each experiment was performed in triplicate using samples from more than three donors. To detect differences between two groups, t-tests were used. The statistical significance of the differences among three or more groups was evaluated using one-way analysis of variance (ANOVA) with Tukey's post hoc tests. All results are presented as means and 95% CIs of the values from three different donors per group.