Hydrostatic pressure promotes chondrogenic differentiation and microvesicle release from human embryonic and bone marrow stem cells

Mechanical stimulation plays in an important role in regulating stem cell differentiation and their release of extracellular vesicles (EVs). In this study, effects of low magnitude hydrostatic pressure (HP) on the chondrogenic differentiation and microvesicle release from human embryonic stem cells (hESCs) and human bone marrow stem cells (hBMSCs) are examined. hESCs were differentiated into chondroprogenitors and then embedded in fibrin gels and subjected to HP (270 kPa, 1 Hz, 5 days per week). hBMSC pellets were differentiated in chondrogenic media and subjected to the same regime. HP significantly enhanced ACAN expression in hESCs. It also led to a significant increase in DNA content, sGAG content and total sGAG/DNA level in hBMSCs. Furthermore, HP significantly increased microvesicle protein content released from both cell types. These results highlight the benefit of HP bioreactor in promoting chondrogenesis and EV production for cartilage tissue engineering.

the procurement of stem cells from the bone marrow is a surgical procedure associated with pain and risk of complications. Also, since these cells are mainly used as autologous cell based therapy, treatments usually involve a two-step operation, leading to higher cost. Furthermore, several studies have reported BMSCs derived from OA patients exhibit reduced chondrogenic capacity. [11] In contrast, there are thousands of human embryonic stem cells (ESCs) lines available from surplus embryos donated from in vitro fertilization procedures and a number of these are of clinical grade. [12] They are pluripotent and possess unlimited self-renewal capacity, thus may form an alternative allogeneic cell source for cartilage tissue engineering applications. The behavior of ESCs has been investigated in a number of naturally derived hydrogels, such as agarose, [13] hyaluronic acid, [14,15] and fibrin gels, [16,17] for cartilage tissue engineering applications.
In vivo, articular cartilage experiences a range of mechanical loading during joint movement, including compression, tension, shear stress, and hydrostatic pressure (HP). [18,19] Mechanical conditioning in vitro has been shown to play a role in chondrogenesis under multiple regimes. [20] The potential for preconditioning of tissue engineered rudiments has been suggested as a way to facilitate engraftment into a repair site in vivo. [20][21][22] How a certain cell type responds to joint-specific mechanical stimulation is therefore crucial to determine its clinical suitability for cartilage tissue engineering application. [4] HP, a key mechanical factor within the joint environment, has been shown to promote matrix synthesis in chondrocytes. [23,24] Studies have also demonstrated cyclic HP upregulated chondrogenic gene expression and/or increased proteoglycan and collagen synthesis in BMSCs, [25][26][27][28] although occasionally conflicting results have also been reported, with HP showing no significant effects. [29,30] However, to the authors' knowledge, to date no studies have investigated the influence of HP on ESC-chondrogenesis. Consequently, there is an urgent need to understand how ESCs respond to HP to further understand how these cells might be used ultimately for clinical therapies.
In addition to cell-based approaches, extracellular vesicles (EVs) are receiving increasing attention as novel acellular tools for cartilage repair. [31,32] EVs are defined as cell-secreted phospholipid nanoparticles which contain a complex biological cargo including nucleic acids, proteins, and other signaling molecules that are believed to stimulate numerous biological processes including proliferation and differentiation. [33,34] In particular, microvesicles (MVs) are a heterogenous population of EVs which are formed by the outward budding of the plasma membrane, possessing a diameter ranging from 100 to 1000 nm. [35][36][37] Several studies have demonstrated the chondroinductive potency of stem cell derived EVs. [31,32] As research continues to foray into the exploitation of EVs in regenerative medicine, there is a need to enhance the scalable manufacture of EVs for clinical applications. [38] Recent studies have shown mechanical stimulation such as shear stress can increase the yield and/or therapeutic potency of EVs secreted by numerous cell types including stem cells. [39,40] It is therefore worth investigating whether HP could provide an alternative approach to enhance the production of EVs for cartilage repair.  [12,41]

Chondrogenic differentiation
hBMSCs were pelleted and differentiated chondrogencially in a highthroughput v-bottomed 96-well plate culture system as previously described. [42] Briefly, 200,000 cells (P3) were added into each well of an autoclave-sterilized v bottomed 96 well polypropylene microplate (Greiner bio-one), and the plate was centrifuged for 5 min at 500× g.
Chondroprogenitors were derived from hESC using a modified version of the directed differentiation protocol described previously. [17] Cells were dissociated and seeded onto VTN-coated tissue culture and changed daily.

Application of HP
hBMSC pellets were subjected to HP from day 1 for a total of three weeks as described previously. [43][44][45] hESCs were first differentiated into chondroprogenitors and then embedded into fibrin gels and subjected to HP for 1 week. For both cell types, HP was applied at an amplitude of 270 kPa at a frequency of 1 Hz, 1 h per day, 5 days per week.
Samples which were not subjected to HP were cultured as parallel controls.

RNA isolation and qRT-PCR
Neo-cartilage tissues engineered from hBMSCs and hESCs (n = 4 per group) were snap frozen upon termination of experiments and homogenized using disposable pellet pestles (Sigma). RNA was then extracted using TRI Reagent (Sigma) and converted into cDNA using High Capacity cDNA Reverse Transcription Kits (Applied Biosystems) (both as per the manufacturer's instructions). Gene expression analysis was performed for SOX9, ACAN, and COL2A1 using SYBR Green-based quantitative real-time polymerase chain reaction (qRT-PCR) with preoptimized QuantiTect primer assays (Qiagen) and an AriaMx Real-Time PCR System (Agilent Technologies). qRT-PCR data were analyzed using the Delta Delta Ct method as described previously [46] with the unloaded control samples used as the calibrator and GAPDH as the endogenous control gene. Relative quantification values are presented as fold changes in gene expression relative to the control group, which was normalized to one.

Detection of EV markers
The presence of EV tetraspanin markers CD9 and CD81 at the surface of EVs was assessed using the ExoViewTM Tetraspanin Kit according to the manufacturers' instructions and as previously described. [48] Briefly IgG spots were used as an isotype control. Data was acquired using the nScan software (Nanoview Biosciences, version 2.8.10) and analyzed using the NanoViewer software (Nanoview Biosciences, version 2.8.10).

Statistics
Statistics were performed using GraphPadPrism software package (San Diego, CA, USA). Unpaired t-test was used to compare between two groups. Statistical significance was considered if p < 0.05. Data are presented as mean ± standard deviation (SD).

HP promotes chondrogenic differentiation of hESCs and hBMSCs
hESCs were first differentiated into chondroprogenitors using a modified version of the directed differentiation protocol described previously. [17] After 14 days of directed differentiation, the hESC marker OCT4 and NANOG significantly reduced whereas the chondrogenic markers including Sox9, COL2A1, and ACAN all increased significantly ( Figure S1A COL2A1 (1.9 fold, p = 0.08, Figure 1). Biochemical data revealed no significant difference in DNA or sGAG content between control and HP group (Figure 2A). sGAG secretion into media was also assayed, however, the level was not detectable (data not shown). There was no significant difference in sGAG/DNA level between the two groups ( Figure 2A). Histologically, more intense Safranin O staining for proteoglycan was observed in the HP group ( Figure 2B), in line with the gene expression data. However, no positive staining for picrosirius red or Type II collagen immunohistochemistry was observed in either group (data not shown).
To study the effects of HP on hBMSCs chondrogenesis, hBMSCs pellets were cultured in a chondrogenic media and subjected to HP (same regime as above) for 3 weeks, while samples that were not subjected to HP were kept as control. HP had no significant effect on expression of any of the chondrogenic genes examined (Figure 3). HP led to a significant increase in DNA (p < 0.01) and sGAG (p < 0.01) content in the samples ( Figure 4A). A significantly higher level of sGAG secretion into the media and total sGAG/DNA (both accumulated in construct and secreted) was also observed with the application of HP (both p < 0.001, Figure 4A). Histologically, more intense Safranin O staining was found in the HP group ( Figure 4B), in line with the F I G U R E 1 Relative SOX9, ACAN, and COL2A1 expression in hESC derived chondroprogenitor seeded fibrin gels that were cultured statically or subjected to cyclic hydrostatic pressure. Gene expression shown relative to GAPDH and data was normalized to unstimulated control. Data were pooled from donor MAN7 and MAN13. *p < 0.05, data are presented as mean ± SD F I G U R E 2 Assessment of hESC derived chondroprogenitor seeded fibrin gels that were cultured statically or subjected to cyclic hydrostatic pressure.

HP increases MV secretion from chondrogencially primed hESCs and hBMSCs
To investigate the effects of HP on hESCs and hBMSCs release of microvesicles (MVs) during chondrogenic differentiation, conditioned F I G U R E 3 Relative expression of SOX9, ACAN, and COL2A1 expression in hBMSCs pellets that were cultured statically or subjected to cyclic hydrostatic pressure. Gene expression shown relative to GAPDH and data was normalized to unstimulated control. Data are presented as mean ± SD media from the loading period was collected and MVs were isolated.  Figure 5D). When MV protein content was normalized to DNA content, HP led to a 1.4-fold increase in MV/DNA level in loaded hESCs (p < 0.001, Figure 5E). For hBMSCs, a small trend of increase in MV/DNA level (1.2 fold) was also observed with the application of HP, although not statistically significant (p = 0.09, Figure 5E).

DISCUSSION
HP is a key mechanical stimulus present in the joint environment.
How progenitor and mature cells respond to HP is crucial to understanding how to achieve clinical success for cartilage repair. In this study, we examined effects of low magnitude HP on chondrogenesis of hESCs and hBMSCs. Here, we showed HP significantly enhanced ACAN expression in chondrogencially differentiated hESCs. It also increased the DNA content, sGAG content and total sGAG/DNA level in tissues engineered using hBMSCs. Furthermore, we investigated effects of low magnitude HP on EV release from cells during the differentiation phase. We found that HP significantly increased MV protein content secreted by both cell types. These results suggest that the use of an HP bioreactor has potential as an effective tool to promote chondrogenesis and EV yield for cartilage tissue engineering application.
Many studies have demonstrated that both physiological level HP (3-10 MPa) and low magnitude HP (100-500 kPa) promotes the chondrogenesis of stem cells. [28] We have previously shown that a low magnitude HP of 270 kPa can promote osteogenesis in bone rudiments and tissue engineered bone. [45,44] Therefore, in this study, we were interested to explore how HP of the same magnitude would affect chondrogenesis of hESCs and hBMSCs. Here, we showed that low magnitude HP resulted in a significant increase in the DNA content, sGAG content, sGAG secretion, total GAG/DNA level as well as the intensity of Safranin O staining of cartilaginous tissue engineered using hBMSCs.
Similar results have also been reported by other studies using low magnitude HP. [27,[49][50][51][52][53] For example, Maxson et al. showed a low magnitude HP of 300 kPa led to increased GAG/DNA level in cartilage tissues engineered using BMSCs. [52] Luo et al. also showed increased DNA content, sGAG content as well as sGAG secretion in BMSCs after being subjected to a HP of 100 kPa for 10 days. [51] While extensive research have been conducted to study effects of HP on chondrocytes, [23,24] BMSCs [4,25] and adipose stem cells, [54] little is known about how ESCs would respond to HP. Here, for the first time, we showed low magnitude HP promoted chondrogenesis of hESCs by inducing a more than 10-fold increase in aggrecan gene expression after application for only rapidly and start to partially dissolve after one week in culture. [55,56] Therefore, in this study we used a relatively short stimulation period for hESCs compared to hBMSCs.
Several studies have reported the influence of physiological conditions such as fluid shear, hypoxia, and oxidative stress on increasing the shedding of MVs from activated cells. [57,58] With the application of HP replicating the in situ environmental conditions within the joint and promoting chondrogenesis in our study, this applied mechanical stimulation could provide a novel approach to promote the scalable manufacture of EVs for cartilage tissue engineering applications.
In order to examine effects of HP on EV yield, we isolated MVs from medium collected during the differentiation phase and characterized F I G U R E 4 Assessment of hBMSCs pellets that were cultured statically or subjected to cyclic hydrostatic pressure. (A) DNA and sGAG content in the sample, sGAG content secreted into medium as well as total sGAG/DNA level (accumulated and secreted) of each group. **p < 0.01, ***p < 0.001, data are presented as mean ± SD. to unstimulated group. [40] The underlying mechanism by which HP regulate stem cell differentiation has not yet been fully elucidated. Previous studies suggest this might happen through HP affecting the endogenous TGF-β production, [27] integrin proteins, [60] intermediate filament, [61] TRP ion channel family, [62] and intracellular calcium stores. [63] A recent review also pointed the possibility of HP affecting primary cilia or nuclei. [28]  In conclusion, this study showed that the application of HP pro-

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.