Time-Dependent Anabolic Response of hMSC-Derived Cartilage Grafts to Hydrostatic Pressure

,


Introduction
Articular cartilage (AC) is a highly specifed tissue that covers the ends of bones within the diarthrodial joints. Its unique biochemical composition is responsible for its load bearing and low friction properties during articulation. Although evolutionary designed to operate within mechanically demanding environments, AC possesses very limited intrinsic regenerative capacity, which can be linked to its avascular, alymphatic, and aneural nature [1]. Additionally, damage and injuries to AC are strongly associated with the development of posttraumatic osteoarthritis (OA) or secondary OA. Tis progressive degenerative disease represents a signifcant societal and fnancial burden worldwide, and the development of an efective therapy represents a great challenge to this day [2].
AC tissue engineering (TE) represents a promising strategy to circumvent some of the shortcomings of conventional therapies for joint regeneration such as autologous chondrocyte implantation (ACI), microfracture, or mosaicplasty [3]. Drawbacks associated with ACI such as donor site morbidity and cell availability could be negated by the use of more readily available undiferentiated mesenchymal stem/stromal cells (MSCs), which in combination with appropriate scafolds and growth factors such as TGF-β3 can be steered to diferentiate into chondrocyte-like cells and deposit cartilage-specifc extracellular matrix (ECM) components [4]. In this context, although the role of growth factors such as TGF-β3 during chondrogenic induction of MSCs in vitro is well established, the efect of removing supraphysiological levels of such factors (e.g., upon implantation in vivo) on engineered graft development remains poorly understood. Terefore, it is paramount to identify conditions that facilitate the development of a stable chondrocyte-like phenotype in chondrogenically primed MSCs. In addition to phenotype maintenance, chondrogenically primed MSCs must respond anabolically to joint-like mechanical loading. However, it remains unclear when TGF-β3-stimulated MSCs adopt such a mechanoresponsive phenotype.
Mechanical stimulation is crucial to skeletogenesis and normal joint development [5]. Interestingly, fracture healing, which is a highly complex regenerative process, is known to recapitulate aspects of embryonic development, where a combination of mechanical forces, specialized cells, and secreted factors drive regeneration of the tissue, albeit the interplay of individual factors is yet to be fully elucidated [6]. Nevertheless, the correct timing of load application has been shown to play a key role during regeneration in various fracture healing models, where delayed application of mechanical load has been shown to be more benefcial in comparison with immediate mechanical stimulation for fracture healing [7,8]. Similarly, for AC regeneration, the in vivo environment appears to orchestrate how progenitor cells diferentiate. For example, MSCs implanted into cartilage defects in mini-pigs have been shown to be less hypertrophic than control cells maintained in vitro [9], pointing to the importance of joint-specifc cues in maintaining a cartilage phenotype. In another study, implantation of prediferentiated human MSCs in a pig model revealed that longer prediferentiation periods in vitro resulted in better collagen type II and sGAG deposition in vivo [10]. Cumulatively, these studies indicate that joint-specifc biomechanical stimuli are integral to directing the diferentiation of stem/progenitor cells and the engineering of functional cartilage grafts; however, the maturity such engineered cartilage tissues needs to achieve to respond anabolically to biophysial cues remains poorly understood.
Hydrostatic pressure (HP) bioreactors have been employed in the studies to emulate the mechanical environment within the joints and to examine how such cues afect the diferentiation of MSCs during chondrogenesis in vitro [11]. It is well established that the appropriate application of HP can beneft chondrogenic diferentiation of MSCs in vitro and facilitates cartilage-specifc ECM deposition, albeit the extent of benefts varies greatly between the experiments [12,13]. Generally, HP has been shown to beneft chondrogenesis when applied during or after various chondrogenic priming periods in the presence of TGF-β3 [11]. It is also possible to use such bioreactors to subject engineered cartilage tissues to joint-like loading conditions, in combination with the removal of TGF-β3 from the culture media, as a model system to better understand how these engineered cartilage grafts will respond to in vivo-like mechanical loads once they are removed from chondrogenic culture conditions [14,15]. We have previously shown that chondrogenically primed human MSCs can respond positively to the application of HP [16]. What remains unclear is how long such engineered grafts need to be maintained in the presence of chondrogenic growth factors before they can mount an anabolic response to the application of HP in the absence of such biochemical cues.
Te goal of this study was to use a HP bioreactor as an in vitro platform to model and understand how chondrogenically primed hMSCs would respond to joint-like mechanical loading in vivo. Specifcally, we sought to explore how the phenotype of hMSCs, which were frst chondrogenically primed for diferent periods of time in the presence of TGF-β3, would change in response to the application of HP. By investigating the anabolic response of hMSCs at diferent timepoints of chondrogenic diferentiation, we aimed to identify the degree of "maturation" necessary for cartilage grafts to potentially respond favourably to the in vivo joint environment. It this way, we hope to demonstrate how such in vitro bioreactor platforms can be utilized for screening of cartilage grafts prior to implantation, which might reduce the number of AC TE grafts that fail in vivo [17].
To this end, hMSCs were cultured for either 7, 21, or 35 days in the presence of 10 ng/ml TGF-β3, at which point this growth factor was removed from the media and the constructs were stimulated with 2 MPa of HP for an additional 7 days. Using this platform, we sought to determine how long hMSCs need to be primed in vitro before they respond anabolically to joint-like loading ( Figure 1). To better mimic the in vivo scenario, where chondrocytes are exposed to oxygen levels between 1-5% in vivo, 5% O 2 tension was employed during cell expansion and diferentiation, conditions which have previously been shown to support more robust chondrogenesis of MSCs [18][19][20][21].

Hydrostatic Pressure Application.
After preculture period in well plates, hydrogels were transferred into gas permeable and watertight cell culture bags (OriGen Biomedical). Heat-sealed bags were then flled either with CDM + without TGF-β3. If not mentioned otherwise, each bag containing 4 ml media/hydrogel was not exchanged during the week of loading to reduce the risk of contamination. For hydrostatic pressure (HP) application, cell culture bags were transferred into a custom-build bioreactor within a 37°C incubator as described previously [15]. Cells were subjected to HP at 2 MPa at 1 Hz for 2 hours daily. Te free swelling controls were placed in the same incubator for the duration of stimulation. After stimulation, all bags were returned to culture incubators (37°C, 5% CO 2 , and 5% O 2 ).

Gene Expression Analysis.
Total RNA extraction was performed using TRIzol ™ reagent (Termo Fisher) in accordance with the manufacturer's instruction. Prior to isolation, samples were physically disrupted with pestles to facilitate equal disruption and dissolution of cells and tissues. After elution and RNA quantifcation (NanoDrop), equal amounts of RNA for each sample were reverse-transcribed to cDNA using High Capacity cDNA Reverse Transcription Kit (Applied Biosystems ™ ). Reverse-transcribed product was quantifed with Qubit ™ ssDNA Assay kit (Termo Fisher), and 10 ng cDNA of each sample was used for real-time PCR. Amplifcation was performed using SYBR ™ Select Master Mix and 7500 Fast System (Applied Biosystems ™ ). B2M functioned as a housekeeping gene and amplifcation of target genes was quantifed using the 2 −ΔΔCt method. KiCqStart ® SYBR ® Green predesigned primers were purchased from Sigma (see Table 1).

Histological Analysis.
Halved samples were fxed in 4% paraformaldehyde overnight at 4°C. Upon fxation, samples were dehydrated in a series of ethanol solutions (50-100%). Subsequently, samples were cleared with xylene and infltrated with parafn wax. All steps from dehydration to infltration were performed in an automated fashion (Leica Biosystems). Samples were then embedded in parafn blocks for microtome slicing (5 µm, Leica Biosystems). All histological stains were performed with an autostainer (Leica). For sGAG determination, alcian blue stain (AB) at pH 1.0 (1% w/v alcian blue 8GX in 0.1 M hydrochloric acid (HCl)) was used; nuclear fast red (0.1% w/v) was used as counterstain. For collagen deposition, picrosirius red (PR) 0.1% w/v was used. All stained slides were imaged with Aperio ScanScope (Leica Biosystems).

2.7.
Immunohistochemistry. Hyaluronidase (Sigma-Aldrich) 1% w/v and pronase (Merck) 3.5 U/ml were used for antigen retrieval (25 minutes at 37°C, respectively). Sections were blocked in the presence of 10% v/v goat serum and 1% w/v BSA (Sigma-Aldrich) at RT and incubated with mouse primary antibodies overnight at 4°C. Sections were blocked in the presence of 10% v/v goat serum and 1% w/v BSA (Sigma-Aldrich) at RT and incubated with mouse primary antibodies against Col10a1 (AB49945, 1 : 200) overnight at 4°C. Goat anti-mouse IgM (AB150121) was used as secondary antibody for Col10a1 sections (4 hours at RT). Goat anti-mouse IgM (AB150121) was used for Col10a1 sections (4 hours at RT); nuclei were counterstained with DAPI (2 μg/ mL, Sigma-Aldrich). All samples were mounted with ProLong ™ Gold Antifade (Invitrogen). Presence of collagen type X was visualized using Leica SP8 scanning confocal microscope. Representative images are shown. Two-photon fuorescence lifetime imaging microscopy (2-P FLIM) was performed on a custom multiphoton system as previously described [23,24]. Briefy, two-photon excitation of nicotinamide adenine dinucleotide phosphate (NAD(P)H) fuorescence was performed with 760 nm excitation wavelength. A 455/90 nm bandpass flter was used to isolate NAD(P)H fuorescence signal. 512 × 512 pixel images were acquired with a pixel dwell time of 3.81 μs and 30-second collection time. At least 3 images for each sample were acquired.

Statistical
Analysis. Statistical analysis was performed using GraphPad (GraphPad Software, USA). All experiments were performed with at least three replicate samples per condition. Te signifcance of mean diferences among two conditions was evaluated using unpaired two-tailed Student's t-test. Te signifcance of diferences among groups was performed using one-way analysis of variance (ANOVA) with a post hoc Tuckey's multiple comparison test. Two-way analysis of variance (ANOVA) with a post hoc Tuckey's multiple comparison test was used to analyse the statistical diferences among groups with two independent variables (e.g., stimulation and duration). Numerical values were presented as mean ± standard deviation. Values of p ≤ 0.05 were considered signifcant.

hMSC Diferentiation and Tissue Maturation over 35 Days in Chondrogenic
Media. Te chondrogenic diferentiation of hMSCs (encapsulated within 2.5% fbrin hydrogels) in response to TGF-β3 stimulation over 35 days in static culture conditions was frst assessed. A continued increase in the expression of the chondrogenic markers ACAN and COL2A1, as well as of the hypertrophic marker COL10A1, was observed over the course of 35 days (Figure 2(a)). No signifcant changes in the expression of the transcriptional factors SOX9 and RUNX2 could be observed after 7 days of culture. A timedependent increase in matrix accumulation was also observed, with sulfated glucosaminoglycans (sGAGs)/DNA levels reaching 81.8 (µg/µg) and collagens/DNA 94.6 (µg/µg) values at day 35 (Figure 2(b)). Histological analysis further confrmed robust chondrogenic diferentiation, with the most  For all examined timepoints, the removal of TGF-β3 did not result in a reduction in the expression of ACAN over the following 7 days of culture; in fact, the expression of ACAN continued to increase after the removal of this growth factor at either day 7 or 21. Te ratio of COL2A1/COL10A1 expression also increased when TGF-β3 was removed from the media after 7 or 35 days of priming. Biochemical assays were used to evaluate how the removal of TGF-β3 infuences overall levels of ECM deposition. No major changes in the overall DNA content over the course of the experiment could be observed (Figure 4(a)). Removal of TGF-β3 did not negatively afect collagen/DNA or sGAG/DNA deposition compared to the timepoint when it was removed (Figures 4(b) and 4(c)). Collagen deposition continued to increase following the removal of TGF-β3 from the media on per cell basis at day 7 or day 21, with similar trends observed for sGAG deposition.  in absence of TGF-β3 was investigated. Upon encapsulation of hMSCs in 2.5% fbrin hydrogels, they were directly mechanically stimulated without any other soluble cues to direct diferentiation. Gene expression analysis of chondrogenic and hypertrophic markers revealed a signifcant upregulation in the expression of the SOX9 transcription factor due to the application of HP ( Figure 5). Despite this, none of the other examined chondrogenic genes were signifcantly afected by the application of HP.

Te Infuence of HP on Chondrogenesis of hMSCs following
Diferent TGF-β3 Priming Periods. Te application of HP did not appear to afect the majority of examined genes following the removal of TGF-β3 from the culture media ( Figure 6). ACAN expression appeared to be positively affected by HP at days 28 and 42, albeit not statistically signifcant. Similarly, no major diferences in expression of key gene ratios could be observed, except for ACAN/COL10A1 ratio that was signifcantly improved at day 42 in the

Normalized to D7
Before TGF-β3 withdrawal 7 days after TGF-β3 withdrawal Before TGF-β3 withdrawal 7 days after TGF-β3 withdrawal Before TGF-β3 withdrawal 7 days after TGF-β3 withdrawal Before TGF-β3 withdrawal 7 days after TGF-β3 withdrawal Before TGF-β3 withdrawal 7 days after TGF-β3 withdrawal Before TGF-β3 withdrawal 7 days after TGF-β3 withdrawal  COL2A1, ACAN, and HDAC4) and hypertrophic (RUNX2 and COL10A1) markers has been normalized to expression levels at day 7 and determined using the 2 −ΔΔCt method. Gene expression ratios for SOX9/RUNX2, COL2A1/COL10A1, and ACAN/COL10A1 were determined using 2 −ΔCt values. Housekeeping gene: B2M. Signifcant diferences are reported using unpaired two-tailed Student's t-test, which was employed to reveal the efect of TGF-β3 at one specifc timeframe. All data were represented as mean ± SD; technical replicates n ≥ 3, * p ≤ 0.05, * * p ≤ 0.01, and * * * p ≤ 0.001.  Figure 4: Biochemical content analysis. Infuence of TGF-β3 priming and subsequent removal on deposition of ECM components. (a) DNA content normalized to wet weight (ww) was determined using Hoechst bisbenzimide H 33258 quantitation assay. (b) Collagen content analysis. Total collagen was determined via hydroxyproline-based quantifcation assay and normalized to wet weight as well as total DNA. (c) sGAG content analysis. sGAG content was determined with DMMB-based assay and normalized to wet weight as well as DNA content. Signifcant diferences are reported using unpaired two-tailed Student's t-test, which was employed to reveal the efect of TGF-β3 at one specifc timeframe. All data were represented as mean ± SD, technical replicates n ≥ 3, * p ≤ 0.05, * * p ≤ 0.01, * * * p ≤ 0.001, and * * * * p ≤ 0.0001.
presence of HP. On the protein level, the application of HP after various priming periods did not appear to have major negative or positive efects on synthesis of ECM components (Figure 7). Of note, the application of HP after 35 days of priming led to a signifcant increase of sGAG/DNA levels at day 42 compared to the free swelling control (Figure 7(c)). Te histological assessment of ECM components deposited over the course of 42 days confrms a very robust chondrogenic diferentiation, as evidenced by the temporal increase of alcian blue and picrosirius red staining at each examined timepoint (Figure 8). Removal of TGF-β3 had no negative efect on the deposition of ECM components at any of the timepoints. Similarly, no discernible diferences between mechanically stimulated and corresponding controls could be observed. Generally, alcian blue stain appeared to be more homogeneously distributed than picrosirius red, which stained weaker at the centre of the constructs at later stages of maturation. Alizarin red staining was performed to examine potential calcium deposition (Supplementary Figure 1); no positive staining could be observed. Successful diferentiation of MSCs is linked to the progressive shift of energy metabolism from glycolysis to oxidative phosphorylation (OxPhos), which can be quantifed by determining protein-bound NAD(P)H levels ( Figure 9). Continuous static culture in the presence of TGF-β3 over 35 days was associated with a gradual increase of average lifetimes of NAD(P)H (τ avg ), indicating a shift towards OxPhos (Figure 2(d)). Te removal of the growth factor after 7 and 35 days of priming resulted in a statistically signifcant increase in τ avg values (Figure 9(a)). Tis efect was further enhanced when HP was applied after 35 days of chondrogenic priming, where τ avg values were signifcantly higher for HP compared to the FS control (Figure 9(b)).

Discussion
As MSCs undergo chondrogenic diferentiation and deposit pericellular and extracellular matrix, their cellular phenotype and associated mechanoresponsiveness might evolve as the function of cell maturity. Several studies have looked at how diferent priming periods afect chondrogenic diferentiation when subsequently subjected to HP; however, employment of diferent cells, 3D environments, and loading regimes made it challenging to develop a generalized statement regarding suitable priming periods [11]. Tis study aimed at assessing how chondrogenically primed hMSCs would respond to hydrostatic pressure following specifc priming periods in the presence of TGF-β3. By culturing hMSCs in the presence of TGF-β3 for prespecifed periods of time and subsequently removing the growth factor prior to application of mechanical stimulation, we also partially mimicked how such cartilage grafts might behave in vivo upon implantation into a mechanically demanding environment. It was found that the removal of TGF-β3 diferentially impacted the development of engineered grafts in a maturity-dependent manner. Furthermore, only mature engineered cartilage grafts responded positively to the application of HP. Tus, an in vitro system for assessing how engineered graft maturity infuences their response to mechanical stimulation was established, which provides  ACAN), hypertrophic (RUNX2 and COL10A1) marker and COLVI and HDAC4 genes were determined using the 2 −ΔCt method. Gene expression ratios for SOX9/RUNX2 and ACAN/COL10A1 were determined using 2 −ΔCt values. Housekeeping gene: B2M. Signifcant diferences are reported using unpaired two-tailed Student's t-test. All data were represented as mean ± SD; technical replicates n � 4, and * * p ≤ 0.01. 8 Journal of Tissue Engineering and Regenerative Medicine valuable information to potentially increase the chances of successful outcomes in vivo. Static culture of hMSCs in 2.5% fbrin hydrogels over the course of 35 days of culture resulted in robust chondrogenic diferentiation, which was refected in the continuous upregulation of chondrogenic genes including COL2A1 and ACAN, as well as accumulation of cartilage-specifc ECM components. To our surprise, no major increase in the expression of SOX9 could be observed after day 7 of culture, although the expression of SOX9 downstream targets such as ACAN and COL2A1 increased continuously until the last examined timepoint at day 35 [26]. Tis observation could potentially be explained by posttranscriptional as well as posttranslational regulation of SOX9 and its downstream targets [27]. Gene expression levels do not reveal information regarding the actual protein levels and their activity. It is possible that translation of SOX9 mRNA is inhibited by appropriate miRNAs, and as maturation of   (SOX9, COL2A1, ACAN, and HDAC4) and hypertrophic (RUNX2 and COL10A1) markers has been normalized to expression levels at day 7. Relative expression levels were determined using the 2 −ΔΔCt method. Gene expression ratios for SOX9/RUNX2, COL2A1/ COL10A1, and ACAN/COL10A1 were determined using 2 −ΔCt values. Housekeeping gene: B2M. Signifcant diferences are reported using two-way analysis of variance (ANOVA) with a post hoc Tukey's multiple comparison test. All data were represented as mean ± SD; technical replicates n ≥ 3, * p ≤ 0.05, * * p ≤ 0.01, * * * p ≤ 0.001, and * * * * p ≤ 0.0001.  Figure 7: Biochemical content analysis. Infuence of the priming period on mechanoresponsiveness of hMSCs. (a) DNA content normalized to wet weight (ww) was determined using Hoechst bisbenzimide H 33258 quantitation assay. (b) Collagen content analysis. Total collagen was determined via hydroxyproline-based quantifcation assay and normalized to total DNA. (c) sGAG content analysis. sGAG content was determined with DMMB-based assay and normalized to total DNA content. Signifcant diferences are reported using two-way analysis of variance (ANOVA) with a post hoc Tukey's multiple comparison test. All data were represented as mean ± SD; technical replicates n ≥ 3, * p ≤ 0.05, * * p ≤ 0.01, * * * p ≤ 0.001, and * * * * p ≤ 0.0001. MSCs progressed, the silencing was progressively attenuated, thus enhancing translation of SOX9 mRNA [28].
TGF-β3 plays a crucial role in various developmental processes including cell diferentiation, proliferation, and ECM deposition [29]. Consequently, TGF-β3 has established itself as one of the most used growth factors in cartilage tissue engineering to promote chondrogenic diferentiation of MSCs [30]. Interestingly, removal of TGF-β3 after 7, 21, and 35 days of priming had no major efects on the expression of SOX9, ACAN, and COL2A1, except for COL2A1 expression at later timepoints (removal of TGF-β3 after 35 days). Tis capacity of short-term exposure to TGFβ to induce robust expression of chondrogenic machinery has previously been demonstrated in specifc contexts. For example, it has been shown that transient short-term exposure of cartilage microtissues, generated from hMSCs, to TGF-β1 was sufcient to induce robust chondrogenic differentiation [31]. Cumulatively, these observations raise the  question whether continuous growth factor supplementation is always necessary for sufcient chondrogenesis and whether fne-tuning of induction kinetics should be leveraged develop novel strategies to achieve hyaline-like phenotype.
Interestingly, the removal of TGF-β3 after 35 days of priming results in a decrease of COL2A1 expression over the next 7 days of culture (compared to day 35), whereas no changes in ACAN are registered. Tere is evidence suggesting that synthesis of collagens in mature chondrocytes is more reliant on the availability of growth factors compared to the synthesis of glycosaminoglycans [32][33][34][35]. Tis might imply that MSCs have acquired a chondrocyte-like state after 35 days of chondrogenic diferentiation. Care must be taken when interpreting the implications of this fnding, as control constructs where TGF-β3 was continuously supplemented into the media over the entire culture period were not included in this study; hence, we cannot be certain whether the same outcome would occur with continuous TGF-β3 supplementation. To our surprise, the expression of COL10A1 appeared to follow the COL2A1 expression pattern. Tis contradicts the reported chondroprotective role of this growth factor, which is also refected in the downregulation of HDAC4 that is responsible for silencing of hypertrophic DNA regions [25,36]. However, TGF-β3 does not signal exclusively through Smad2/3 and has been shown to phosphorylate Smad1/5/8 as well, which is associated with upregulation of hypertrophic and osteogenic markers. Such hypertrophic diferentiation represents one of the main challenges of articular cartilage TE using MSCs [37,38]; therefore, further work is required to identify how specifc growth factor priming regimes infuences progression along an endochondral pathway. Nevertheless, removal of TGF-β3 did not have a negative impact on chondrogenic/hypertrophic gene ratios, suggesting that removal of TGF-β3 does not necessarily lead to progression along an endochondral pathway; this was also supported by the relatively weak staining for collagen type X and calcium deposition (Supplementary Figures 1 and 3).
Although the application of HP is generally benefcial during chondrogenic diferentiation of MSCs, it is not known whether HP is sufcient to induce diferentiation in the absence of exogenous growth factors [11,39,40]. Application of HP alone for one week resulted in a statistically signifcant upregulation of SOX9 compared to the FS control. However, this upregulation was signifcantly lower (3.5-fold) compared to the growth factor induced increase (40-fold) after one week of culture ( Supplementary Figure 2). It is therefore questionable whether HP induced upregulation of SOX9 would be sufcient to initiate expression of its downstream targets.
It is well accepted that application of mechanical cues such as hydrostatic pressure can beneft tissue engineering of hyaline-like cartilage [13,14]. Interestingly, the extent of the beneft appears to be contextual and dependent on various factors including cell source, culture format, biomaterial, and mode of stimulation [11,12]. Tis challenges the identifcation of the most appropriate in vitro bioreactor conditions to enhance diferentiation and ECM deposition. For example, previous studies from our lab have shown that porcine MSCs embedded within agarose and fbrin hydrogels respond positively to the application of HP [14,41]. However, when human MSCs were encapsulated into fbrin hydrogels, no benefts of HP application could be identifed [16,42]. One possible explanation for such a discrepancy could be the employment of diferent HP magnitudes (10 vs 2 MPa). However, lower magnitudes have been previously shown to support chondrogenesis in a collagen scafold and in pellet cultures employing human MSCs [39,[42][43][44]. Tis points to intricate interdependencies that need to be identifed and leveraged to enable the engineering of hyaline-like cartilage using stem/progenitor cell populations. In this context, there is also very little knowledge regarding the impact of HP on hMSC fate after specifc chondrogenic priming periods [14,15,45,46]. Based on the quantitative gene expression data, neither negative nor positive efects of HP application could be observed for most of the examined genes at any timepoints. ACAN was the only gene that appeared to respond positively to mechanical stimulation after 35 days of priming. Tis observation was further confrmed by the biochemical content analysis that revealed signifcant increase of sGAG/DNA at these later timepoints. An evaluation of chondrogenic/hypertrophic gene ratios also indicates that mechanical stimulation is supportive of the chondrogenic phenotype after longer priming periods. Cumulatively, this data suggests that priming/culture periods and the "maturity" of chondrogenically primed hMSCs can play a considerable role in how mechanical cues are perceived by diferentiating cells. Tese results are in line with observations made in vivo, where delayed application of stress was most benefcial in terms of regeneration [7,8,10].
It has been reported that maturation of MSCs is linked to metabolic changes, where the glycolytic mode of energy generation shifts towards oxidative phosphorylation (OxPhos) [47]. We could confrm the incremental increase of protein-bound NAD(P)H levels over the course of 35 days in static culture using FLIM [48]. Although mature chondrocytes are known to utilize glycolysis in their hypoxic niche for energy generation, this glycolytic state appears to be more pronounced in naïve MSCs [49]. Interestingly, the removal of TGF-β3 appeared to facilitate the shift towards OxPhos at all examined timepoints. Tis indicates that chondrogenesis-associated metabolic shifts can be induced by short-term exposure to growth factors and should be considered during the optimization of chondrogenic culture protocols. More importantly, the application of HP at later timepoints appeared to further progress the metabolic shift towards OxPhos, which has been associated with maturation of MSCs and has been shown to reduce downstream functional features of osteoarthritis [47,50]. In summary, this points to the essential role of mechanical cues in the maintenance of the chondrogenic phenotype in diferentiating MSCs in a maturitydependent manner.

Conclusion
Herein we have established an experimental platform that partially mimics the transition of a cartilage graft from an in vitro culture into a mechanically stimulated in vivo environment. Tis was achieved by culturing fbrin embedded hMSCs in chondrogenic media with TGF-β3 and subsequently subjecting these constructs to hydrostatic pressure following the removal of exogenous growth factors. By utilizing 1-, 3-, and 5-week priming periods, we have examined how the maturation state of chondrogenically primed MSCs afects their response to HP. Subjecting cells to HP was only benefcial at later stages of diferentiation, which was refected by an upregulation of ACAN, sGAG deposition, and metabolic activity upon mechanical stimulation. Tese results suggest that in this specifc experimental setup a priming period of at least 5 weeks would be advisable before in vivo implantation of such an engineered cartilage graft.

Data Availability
Te data used in this study are available on request from the author.

Ethical Approval
Te authors confrm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to. No ethical approval was required.

Conflicts of Interest
Te authors declare that they have no conficts of interest. counterstained with DAPI (2 μg/mL, Sigma-Aldrich). Presence of collagen type X was visualized using Leica SP8 scanning confocal microscope. (Supplementary Materials)