Spinning around or stagnation - what do osteoblasts and chondroblasts really like?
© I. Holzapfel Publishers 2010
Received: 3 April 2009
Accepted: 5 August 2009
Published: 29 January 2010
Skip to main content
© I. Holzapfel Publishers 2010
Received: 3 April 2009
Accepted: 5 August 2009
Published: 29 January 2010
The influcence of cytomechanical forces in cellular migration, proliferation and differentation of mesenchymal stem cells (MSCs) is still poorly understood in detail.
Human MSCs were isolated and cultivated onto the surface of a 3 × 3 mm porcine collagen I/III carrier. After incubation, cell cultures were transfered to the different cutures systems: regular static tissue flasks (group I), spinner flasks (group II) and rotating wall vessels (group III). Following standard protocols cells were stimulated lineage specific towards the osteogenic and chondrogenic lines. To evaluate the effects of applied cytomechanical forces towards cellular differentiation distinct parameters were measured (morphology, antigen and antigen expression) after a total cultivation period of 21 days in vitro.
Depending on the cultivation technique we found significant differences in both gen and protein expression.
Cytomechanical forces with rotational components strongly influence the osteogenic and chondrogenic differentiation.
Bone grafting is a common procedure in orthopaedic surgery and the implantation of autologous bone grafts supplying osteoinductive growth factors, osteogenic cells, and a structural scaffold, has become the gold standard for the surgical treatment of bone defects caused by trauma, tumor, infection or congenital abnormalities. In addition, bone grafts are frequently used for spinal fusion, joint revision surgery, corrective osteotomy and bone reconstruction. The amount of bone available for autografting is limited and bone graft harvesting procedures are associated with a multitude of risks, such as pain, neurovasculare injury, persisting haematoma or infection at the donor site [1–3]. The application of allograft bone as an alternative treatment option carries the potential risk of infection and graft failure as a consequence of the reduced osteoinducitvity of allograft bone . Several biomaterials such as metal alloys, ceramics or bone cements have been used for decades as permanent implants to overbridge or stabilize bone defects. Although those bone substitutes have proven utility, they have often resulted in complications such as stress shielding-induced resorption of the surrounding bone and fatigue failure of the implant.
During the last years tissue engineering based treatment concepts and cell therapeutics showed promising results in vitro. Mesenchymal stem cells (MSCs) can easily be isolated and expanded from bone marrow (BM) aspirates. Because of their capacity for ex vivo proliferation and differentiation they provide a good source of osteoprogenitor cells within custom-shaped scaffolds for implantable autologous bone tissue thus allowing the generation of a large transplantable cell population from a small biopsy [5–11].
However, the influcence of sheer stress in cellular migration, proliferation and differentation of MSCs is still poorly understood in detail. Most experimental designs consider laminar or rotation flow, dynamic or hydrostatic pressure, and bending or compressive strain devices to evaluate cytomechanical in vitro-effects. One limitation of the static cultivating technique is the inhomogenous oxygen and nutrient concentration and transport within the cellular carrier (scaffold), resulting in a decrease of differentiation and proliferation an thus restricting the size of the scaffolds [9, 12]. Different bioreactor systems have been used to overcome such limitations, mimicking certain aspects of the native cell environment of functional tissues and providing physiologically relevant physical signals [13–15]. Recent investigations have shown that spinner flasks applied in cell culture to regenerate cartilage and bone tissue can improve cellular distribution and differentiation in scaffolds [16–19].
For the quantification of cellular differentiation at the molecular level, osteogenic differentiation of MSCs is controlled by the interaction of hormones and transcription factors: runt-related transcription factor-2 (RUNX2) effectuates the expression of bone-specific genes, e.g. osterix (OSX), collagen type 1 alpha-1 (COL1A1), osteocalcin (OC), and bone sialoprotein (BSP) by binding to the promoters of these genes. Generally, alkaline phosphatase (ALP), COL1A1, BSP, RUNX2, transforming growth factor-beta 1 (TGFB1), osteonectin (ON), and bone morphogenetic protein-2 (BMP2) are known to be early markers of osteoblastic differentiation, whereas OC and osteopontin (OPN) are expressed later in the differentiation process .
In the presented study, the MSC cells were cultured in either osteogenic or chondrogenic induction medium and then incubated for 21 days into three culture system designs, including static culture (group I, STAT), spinner flask bioreactor (group II, SPUN) and rotating wall vessel reactor (group III, RWV). The aim of our study was to investigate and compare gene and protein expression after different cytomechanical forces were applied.
Human MSCs were isolated via density gradient centrifugation after bone marrow aspiration from a healthy, 39-year old female donor volunteer after informed consent was obtained according to the Declaration of Helsinki in its present form. Cell culture conditions were DMEM-low glucose media, 20% fetal calf serum (FCS)-gold (all agents PAA Laboratories, Cölbe, Germany), L-glutamin with 1% penicillin/streptomycin (PAA) in tissue culture polystyrene flasks in 5 Vol. % CO2 at 37°C. Medium was exchanged twice a week. Adherent cells, judged 80% - 90% confluent by phase contrast microscopy, were detached mechanically and passaged supported by 0.05% trypsin/0.02% EDTA solution (PAA). Cells from 3rd passage were re-suspended and a total number of 3 × 106 cells were cultivated onto the surface of a 3 × 3 mm porcine collagen I/III carrier (BioGide®, Geistlich Pharma AG, Wolhusen, Switzerland) which consists of a porous, porcine-derived, semipermeable, resorbable, non-crosslinked porous collagen I/III membrane (height 1000 μm). According to the producer's information, telomer peptides were removed during manufacture .
After an incubation time of 30 minutes to allow cellular adherence, cell cultures were transfered to the different cultures systems: regular tissue flasks (group I, 50 ml), spinner flasks (group II, 100 ml, 93 UpM) and rotating wall vessels (group III, 50 ml, 25 UpM). Following standard protocols  cells were stimulated lineage specific towards the osteogenic [(DMEM-low glucose (PAA), dexamethasone (Sigma Taufkirchen, Germany), L-ascorbic-2-phosphate (Sigma), β-glycerolphosphatate (Sigma)] and chondrogenic [(DMEM-high glucose PAA), insulin/trans ferrin/selenic acid (ITS) (Sigma), dexamethasone (Sigma), Lascorbic-2-phosphate (Sigma), TGF-β1 (Sigma), pyruvate (Sigma)] lines. After a total cultivation period of 21 days in vitro the following techniques were applied and parameters were measured for evaluation.
For immunocytochemical evaluation, we used 10 μm frozen sections (Cryostat, Zeiss, Göttingen, Germany) that were fixed onto slides (Superfrost plus, Menzel, Braunschweig, Germany) after an in vitro follow-up of 21 days. To detect osteogenic differentiation we used anti-collagen I, anti-collagen III (Chemicon Int., Hampshire, UK), anti-osteocalcin (OC) (Santa Cruz Europe, Heidelberg, Germany) alkaline phosphatase (ALP; Blue Alkaline Phosphatase Substrate Kit III, #SK-5300, Vector Laboratories, Burlingame, CA), anti-RANKL (Santa Cruz), CD34, CD105 (CD34 and CD105 antibodies; DAKO-Cytomation, Hamburg, Germany). For the detection of chondrogenic differentiation we used anti-collagen II (Chemicon Int., Hampshire, UK), chondrogenic oligometric matrix protein (COMP) (AbD Serotec, Düsseldorf Germany) cartilage proteoglycan (CP) (Chemicon), anti-Runx2 (R&D Systmes, Wiesbaden, Germany). In addition, the mesenchymal marker CD105 and the hematopoietic antigen CD34 were analyzed.
The polymer samples were fixed in 4% paraformaldehyde (Roth, Karlsruhe, Germany) at 20°C for 30 min and rinsed in PBS. Endogenous peroxidases of the specimen were blocked by 0.3% perhydrol-PBS solution. After rinsing in PBS, the cell culture dishes were incubated with primary antibodies against different antigens for chondrogenic (CD34, CD105, collagen II, COMP, CP, Runx2) und for osteogenic (CD 34, CD 105, collagen I, collagen III, OC, Rankl) with further incubation at 4°C for 12 h. A second antibody system (Anti-mouse-IgG biotinylated, anti-rat-IgG biotinylated, anti-Rabbit-IgG biot., anti-goat-IgG biot. and avidin-biotin complex (all Vector) and 3,3-diaminobenzidine (Sigma) was used for optical visualization. ALP activity was measured after direct substrate incubation (SK-5.200; Vector) for 30 min at RT in the dark. The cell cultures were analyzed and blinded by an independent observer using episcopic light microscopy (Axiovert 200; Zeiss) in combination with a computer-supported imaging picture analysis system (Axiovision; Zeiss). Feulgen staining (Feulgen-Kit; VWR, Darmstadt, Germany) served for semiquantitative DNA evaluation. Specimens were air-dried for 1 h, fixed in buffered 4% paraformaldehyde (pH 7.0), rinsed in water for 10 min, and incubated in 5 M HCl at 22°C for 50 min. Specimens were then washed twice in aqua dest., transferred to Schiff 's reagent for 60 min at RT, rinsed in bisulfite solution twice (3 min) and in aqua dest. 2 min at RT, followed by dehydration in graded alcohols.
At follow-up, bone marrow cells were morphologically analyzed using phase-contrast microscopy (Axiovert 200, Zeiss) supported by a computer picture analysis system (Axiovision, Zeiss). For semiquantitative analysis of the antigen expression by polystyrene adherent cells, the following score was used: no cells: 0; single cells (< 10% of the surface): 1; sub-confluent monolayer of + cells (10%-59% of the surface): 2; confluent monolayer of + cells (> 60% of the surface): 3.
For mRNA analysis, the adherent cells were removed from culture dishes supported by 0.05% trypsin/0.02% EDTA solution (PAA)and resuspended in each 350 ml RLT buffer (Quiagen, Hilden, Germany) supplemented by 1: 100 14.3 M beta-mercaptoethanol (VWR, Darmstadt, Germany). One-step RT-PCR was performed using a thermal cycler (Mastercycler Gradient, Eppendorf AG, Hamburg, Germany). The reaction mixture (25 μL) contains 2 μl of human RNA, 1 μL 10 mM dNTP-Mix (Qiagen)), 1 μL recombinant RnasinTM/ribonuclease inhibitor 1:4 (Promega, Madi son, WI), 1 μl of each primer, 1 μl enzyme mix, 5 μl RT-buffer, 5 μl Q-solution, and 8 μl RNAse free aqua (Quiagen). For transcription and amplification, we used an enzyme mix containing OmnicriptTM Reverse transcriptase, SensiciriptTM Reverse Transcriptase, and HotStarTaq DNA polymerase (OneStep RT-PCR Kit 100, Quiagen). The thermal cycle conditions used were as follows: 30 min at 50°C (reverse transcription), 15 min at 95°C (denaturation) followed by RT-PCR cycling of 95°C for 30 s (denaturation), 55°C for 30 s (annealing), and 72°C for 60 s (extension), and lastly a final extension for 60 s at 72°C. RT-PCR was performed 35 cycles. GAPDH used as housekeeping gene. RT-PCR products were combined and resolved in a 2% agarose gel stained with ethidium bromide. Documentation and semiquantitative evaluation was performed using a gel documentation system and software (including a real-time camera with ethidium bromide filter combined with Alpha-DigiDoc TM RT software, Alpha Innotec, San Leonardo, CA).
Osteocalcin (OC): 50-agt cca gca aag gtg cag c-30/30-ggc cgt aga gcg ccg at- 50 (NM_199173),
RANKL: 50-cag agc gca gat gga tcc t-30/30-gta cca aga gga cag act ca- 50 (NM_003701),
Runx2: 50-cct tca agg tgg tag ccc t-30/30-gcc tgg ggt ctg taa tct g-50 (NM_004348),
Collagen 1A1: 50-gat ggc tgc acg agt cac a-30/30- gcc aga tgg caa ggc ttc t-50 (NM_000088),
Collagen 2A1: 50-ttt ccc agg tca aga tgg tc-30/30-ctg cag cac ctg tct cac ca-50 (NM_001844),
Collagen 3A1: 50- gga cat cga gga ttc cct g-30/30- cac tct tga gtt cag gat gg-50 (NM_000090),
GAPDH: 50-ctc aag atc atc agc aat gcc-30/30-gat ggt aca tga caa ggt gc- 50 (NM_002046),
COMP: 50- ctg gct gtg ggt tac act g-30/30- gat gat gtt ctc ctg gga g-50 (NM_000095),
CD 34: 50- cat cac aga aac gac agt caa-30/30- ctg cct tga tgt cac tta gg-50 (NM_001773.1),
CD 105: 50- ggc cgc acg ctc gag tg-30/30- aca tga gca gct ccg ggc-50 (NM 000118).
Depending on the cultivation technique we found significant differences in both, gen and protein expression. The semiquantitative antigen pattern before lineage stimulation was:
ALP (+++), col I (+++), col II (-), col III (-), COMP (+++), CP (-), RANKL (++), Runx2 (+), OC (+), CD34 (-), CD105 (+++).
In our study we showed that cytomechanical forces with rotational components influence the osteogenic and chondrogenic in vitro-differentiation of MSCs. In contrast to the static forces or convential tissue flasks, spinning incubators are characterized by sheer forces, which may have different effects on cellular proliferation, differentiation and protein expression on bone marrow cells. The finding that sheer forces promote osteoblastic differentiation is corresponding to the current hypothesis suggesting that interstitial fluid flow is an important component to stimulate putative mechanosensors such as integrins in bone cells. For example, it was shown that activated integrins transduce their signal to the actin cytoskeleton and induce changes in the gene and protein expression such as cyclooxytgenase 2, cfos or NO . In addition, there is evidence that also connexin 43 hemichannels represent another distinct subgroup of mechanosensors which are activated by stretch forces but also by fluid flow allowing paracine factors (small signalling molecules) to induce intracellular signalling pathways [29–33]. It was demonstrated by McAllister et al.  that fluid shear stress is the determining stimulus in the response of osteoblasts. Our results that spinning cultures promote osteoblast differentiation stronger than a rotating wall vessel is also corresponding to other authors who showed that the formation of actin stress fibers in osteoblasts lasts 1 h of steay fluid flow compared to 5 h of oscillatory fluid flow . As demonstrated by Li, boths oscillatory and also steady fluid flow enhance expression of osteoblastic markers such as osteopontin and osteoclacin and are related to PGE2 release .
Also the osteoblast promoting effects induced by dynamic lodings and oscillating fluid flow are reproduced by the tubulences within the scaffold in a spinning incubator.
Our results indicate that cytomechanical sheer forces and medium flow induce cellular proliferation. The spinning culture system and static forces (control) seem to promote osteoblast differentiation. In contrast, both chondrogenic markers (COMP, CP) were expressed only by those cell cultures which were exposed to the RWV system. However the different number of antigen positive cells on the collagen I/III carrier can also be based on different cellular proliferation rates in the defined systems.
It has been demonstrated on MSC that fluid flow can induce a significant increase in proliferation within one hour via the activation of MAP and calcium signalling cascades .
Besides direct cytomechanical effects on cellular gene and protein expression the application of bioreactor systems also promotes the supply and local concentration of nutrients into porous scaffolds.
The transport of nutrients and waste products within a cell-seeded matrix derives primarily from diffusion . In the absence of fluid flow, the primary mechanism allowing transport of nutrients to the center of scaffolds cultured under static conditions is diffusion, which may not meet the significant metabolic requirements of cells seeded on three dimensional scaffolds and cultured for extended time periods . This may explain the cell proliferation rate in static culture was not comparable to dynamic spinner flask and rotating wall vessel culture conditions. Two different bioreactor culture systems were performed in parallel to the static culture conditions. These bioreactor culture systems were used to compare our dynamic cultures with established static culture approaches.
The static culture represents one established methodology for the culture of BM-derived adhesion-dependent cells. In static culture, MSCs were cultured without exposure to hydrodynamic shear forces, and with mass transfer by molecular diffusion only. In spinner flask culture, flow and mixing of culture medium was associated with turbulent shear at constant surfaces. Mass transport between the constructs and culture medium was enhanced by convection, whereas mass transport within the constructs remained controlled by diffusion. In rotating wall vessel culture, the utility provides a microgravity environment with low fluid shear and dynamic intercellular interactions [14, 40]. It is suggested that shear stress is an important biomechanical parameter in regulating human MSC construct development .
These MSCs are cultivated under different culture systems with same osteogenic induction medium. We found that some proteins as ALP mentioned earlier were expressed in spinner dynamic culture when compared with static culture condition, which means that mechanical stress had effects during osteogenesis process. Some studies indicate that intermittent mechanical loading could promote chondrogenesis of BM-derived MSCs through responsive gene regulating pathway [42, 43]. On the other hand, the mechanical stimulation was minimized because of the microgravitational environment in the rotating reactor culture system. The expression of osteoblastic and chondrocytic lineage proteins appeared to be obviously inhibited.
A significant difference in gene expression between the different culture conditions was observed. The better mixing provided in the spinner flasks may explain the accelerated proliferation and differentiation of MSCs, and the localization of the enhanced mineralization on the external surface of the scaffolds. The results were in agreement with the previous research studies reported by Sikavitsas et al  and Wang et al . Wang et at concluded, that the mechanically active environment present in the spinner flask bioreactors mediates the effectiveness of adult human MSCs in 3D scaffolds for osteogenic and chondrogenic differentiation .
The trends observed in the cell/polymer constructs cultured in the spinner flask bioreactor and the static culture are in agreement with the temporal expression and cell growth of osteoblastic cells described. Similar cellular behavior was seen by Goldstein et al.  for the early osteoblastic differentiation.
MSCs represent a promising, readily available autologous cell source for bone tissue-engineering applications [26, 27, 44–48]. Several preclinical studies directed at de novo osteogenesis have been conducted using rat marrow stromal cell culture [17, 18, 49–53] or osteoblastic cell lines [24, 54, 55] as a surrogate for primary human cells. However, primary human MSCs are the idal cell type for such a study because of interspecies variation.
To investigate the efficiency of bone tissue-engineering protocols using primary human MSCs, researchers would ideally include replicates of primary cell cultures from a sufficient number of donors to account for biological variation. The data obtained should then be stratified by relevant factors, e.g. sex, age, and comorbidity. However, such a setup seems far from realistic due to limitations in experimental replicates and restricted availability of primary human MSCs cultures. Another problem is that primary MSCs gradually lose both their proliferation and differentiation potential during ex vivo expansion . Thus early-passage cultures are needed for preclinical studies, further limiting the amounts of biological material available. MSC cultures from young patients are of particular interest because especially young patients may have musculoskeletal diseases and would benefit from the implantation of durable, tissue-engineered, autologous bone graft composites. In the present study, we used a well-characterized immortalized human BM-derived MSC population.
Several studies show that fluid shear stress upregulates anabolic factors relating to proliferation and differentiation of adherent bone cells [56–58]. The effect of fluid flow on cellular response is known to involve typical mechanotransduction pathways, e.g. integrin activation, influx of extracellular Ca2+, and activation of transcription factors via second messenger systems resulting in the upregulation of bone-related genes and the expression of signaling molecules . Fluid flow stimulation of ex vivo 3-D cultured osteoprogenitor cells therefore represents a promising approach to mimic the native environment of these cells in vivo.
To combine fluid flow stimulation and mitigation of external mass transport limitation, a number of dynamic culture strategies have been developed. Among those systems, bioreactors based on axial perfusion of the cell-loaded scaffold [17, 19, 49, 52, 53, 60] and spinner flask bioreactors [17–19] have been tested for bone tissue engineering. The spinner flask is recognized as a controllable system for minimizing diffusional gradients within cell/scaffold constructs and exposing 3-D cultured cells to fluid convection by mixing of culture medium . Here, we investigated the efficiency of this type of bioreactor system in relation to proliferation, distribution, and osteogenic differentiation of 3-D cultured human MSCs.
Cellularity of the scaffolds was assessed by total DNA assay. The results generally indicated a continued proliferation of human MSCs within the scaffolds for up to 3 weeks of culture, regardless of the culture method. Sikavitsas et al. reported that continuous fluid convection significantly stimulated proliferation of rat marrow stromal cells after 7 and 14 days, but not after 21 days of 3-D spinner flask culture compared with static culture . In a similar study, Goldstein et al. found the same tendency for 7 and 14 days periods of culture .
To come back to the initial question wheter osteoblasts prefer spinning around or stagnation we assume that stagnation (static load, gravity) but also spinning allow for osteoblast in vitro differentiation. In contrast chondroblasts prefer continuus sheer forces as simulated by the rotation vessel. However, a different situation was found in scaffold-associated cultures. Here boths, the STAT and in less amounts the RWV promote osteogenic differentiation. The same tendency was found for chondroblast differentiation under chondrogenic mixture. The absence of col II can be interpretated as a wash-out effect promoted by short molecule length compared the relatively large collagen I triple-helix . The polarity in antigen-expression dependent on the semipermeable or the permeable (highly porous) site of the scaffold reflects the strong influence of surface parameters on cell differentiation. The dynamic medium flow with turbulences does not allow any conclusion of the effects of further surface geometry paramters such as convexity or concavity in our investigation.
We used MSCs derived from adult human BM and applied three different culture environments (static culture, spinner flask, rotating wall vessel) to study osteogenesis and chondrogenesis under controlled in vitro conditions. Our results are in agreement with other studies and suggest that osteogenesis and chondrogenesis in cultured MSCs can be modulated by fluid flow and cytomechanical forces. Here, spinner flask bioreactor systems can mimic some aspects of the native biophysical environment and could be utilized in controlled studies on cell function and tissue engineering.