Edited by: Martin James Stoddart, AO Research Institute, Switzerland
Reviewed by: Kar Wey Yong, University of Calgary, Canada; Luca Pierelli, Sapienza University of Rome, Italy
This article was submitted to Tissue Engineering and Regenerative Medicine, a section of the journal Frontiers in Bioengineering and Biotechnology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
The mechanobiological behavior of mesenchymal stem cells (MSCs) in two- (2D) or three-dimensional (3D) cultures relies on the formation of actin filaments which occur as stress fibers and depends on mitochondrial dynamics involving vimentin intermediate filaments. Here we investigate whether human platelet lysate (HPL), that can potentially replace fetal bovine serum for clinical-scale expansion of functional cells, can modulate the stress fiber formation, alter mitochondrial morphology, change membrane elasticity and modulate immune regulatory molecules IDO and GARP in amnion derived MSCs. We can provide evidence that culture supplementation with HPL led to a reduction of stress fiber formation in 2D cultured MSCs compared to a conventional growth medium (MSCGM). 3D MSC cultures, in contrast, showed decreased actin concentrations independent of HPL supplementation. When stress fibers were further segregated by their binding to focal adhesions, a reduction in ventral stress fibers was observed in response to HPL in 2D cultured MSCs, while the length of the individual ventral stress fibers increased. Dorsal stress fibers or transverse arcs were not affected. Interestingly, ventral stress fiber formation did not correlate with membrane elasticity. 2D cultured MSCs did not show differences in the Young's modulus when propagated in the presence of HPL and further cultivation to passage 3 also had no effect on membrane elasticity. In addition, HPL reduced the mitochondrial mass of 2D cultured MSCs while the mitochondrial mass in 3D cultured MSCs was low initially. When mitochondria were segregated into punctuate, rods and networks, a cultivation-induced increase in punctuate and network mitochondria was observed in 2D cultured MSCs of passage 3. Finally, mRNA and protein expression of the immunomodulatory molecule IDO relied on stimulation of 2D culture MSCs with pro-inflammatory cytokines IFN-γ and TNF-α with no effect upon HPL supplementation. GARP mRNA and surface expression was constitutively expressed and did not respond to HPL supplementation or stimulation with IFN-γ and TNF-α. In conclusion, we can say that MSCs cultivated in 2D and 3D are sensitive to medium supplementation with HPL with changes in actin filament formation, mitochondrial dynamics and membrane elasticity that can have an impact on the immunomodulatory function of MSCs.
For clinical-scale expansion of functional mesenchymal stem cells (MSCs) the use of xeno-based serum products are prohibited, but human platelet lysate (HPL) can potentially replace fetal bovine serum (FBS) the most widely used medium supplement in the past (Henschler et al.,
An important issue for cultivation and expansion of MSCs, next to the culture medium selection, is the mechanobiological effect of 2D cultivation (the adherence of MSCs to coated plastic surfaces) or 3D cultivation (MSCs grown in spherical aggregate where cells produce autologous matrix components) (Dominici et al.,
Due to the potential of MSC application in regenerative medicine, the role of mitochondrial metabolism during cultivation in different culture media, have received substantial attention from life scientists and clinicians (Hsu et al.,
Here we investigated the actin filament formation and mitochondrial morphology in 2D and 3D MSC cultures by flow cytometry or laser scanning confocal microscopy (LSM) in response to HPL, applied an MSC migration assay, studied force-distance profiles by atomic force microscopy (AFM) and analyzed the expression of immunomodulatory molecules GARP or IDO following pro-inflammatory stimulation with a combination of IFN-γ and TNF-α by RT-PCR and western blotting (Probst-Kepper and Buer,
The study was approved by the ethic commission of the Medical University Vienna (EK791/2008, EK1192/2015), the University Hospital of Lower Austria (GS1-EK-4/312-2015) and the Danube University Krems (Nr. 821/2009). The placenta was obtained from healthy delivering woman in accordance with the Austrian Hospital Act (KAG 1982) after written informed consent was signed. The amnion was separated from the placenta and cut in 3 × 3 cm squares. The slices were washed in physiological NaCl and digested with dispase (2.5 CU/ml, Becton Dickinson, Franklin Lakes, NJ) for 9 min at 37°C (Soncini et al.,
Adherent amnion derived stromal cells in passage 0 (P0) were grown to 80% confluence and the monolayers rinsed with PBS. Accutase (Gibco, Thermo Fisher Scientific) was added, incubated at 37°C until cells detach and stopped by adding 5 ml of PBS. The cells were centrifuged at 300 g for 5 min and counted. Five thousand cells/25 μl drop were pipetted on a lid from a 100 mm petri dish (Greiner Bio-One, Kremsmünster, Austria) and incubated as hanging drops at 37°C in 5% CO2 humidified environment. The aggregate formation was monitored by a stereo microscope (Olympus IMT2) and spheroids were collected after 2 days and forwarded to characterization by flow cytometry.
Living MSCs (Live/Dead Cell Assay, Invitrogen, Thermo Fisher Scientific) from 2D cultures or spheroids in P1 were characterized for the expression of ecto-5'-nucleotidase (APC CD73), Thy-1 a glycophosphatidyl-inositol (GPI) anchored conserved cell surface protein (FITC CD90), endoglin a component of the receptor complex of TGF-β (PE-Cy7 CD105) and for glycoprotein A repetitions predominant (eFluor660® GARP) mAb (1 μg/ml, all from eBioscience, Thermo Fisher Scientific) on a Gallios 10/3 flow cytometer (Beckman Coulter GmbH, Krefeld, Germany). Actin and mitochondria quantification was performed using a AlexaFluor® 594 labeled phalloidin (1 U/ml) or MitoTracker™ Green FM (100 nM, both from Molecular Probes, Thermo Fisher Scientific) and size measurements were performed with silica beads [1.5 μm (Kisker Biotech, Steinfurt, Germany), 30 and 65 μm (Beckman Coulter GmbH)].
Spherical MSC aggregates or a suspension of 1 × 106 MSCs/ml were incubated in 24 well plates (Nunc, Thermo Fisher Scientific) on Nunc™Thermanox™ coverslips (Nunc, Thermo Fisher Scientific) for 16 h at 37°C in 5% CO2 humidified environment. After three times washing with PBS, the coverslips were fixed with 2.5% glutaraldehyde (Sigma-Aldrich, St. Louis, MO) for 20 min at 4°C, washed two times again and dehydrated by stepwise alcohol extraction from 50 to 100%. The coverslips were mounted on EM-Tec CT12 conductive double side adhesive carbon tabs (Micro to Nano V.O.F., Haarlem, Netherlands) with a diameter of 12 mm and then surface sputtered with gold using a QuorumTech Q150T ES (Quorum Tech Ltd., Laughton, UK) for better resolution. For the scanning electron micrographs of spherical MSC aggregates and single adherent MSCs a FlexSEM 1,000 scanning electron microscope (Hitachi Ltd. Corp., Tokyo, Japan) with SEM MAP camera in SEM mode was used.
MSCs in P1 in 2D or 3D cultures were stimulated with 50 ng/ml TNF-α and 50 ng/ml 313 IFN-γ (both from PeproTech, Rocky Hill, NJ) in MSCGM™- 314 medium or MSCBM™ with 8% HPL for 24 and 48 h.
MSCs were grown in chamber slides for 72 h or alternatively MSC spherical aggregates were adhered to plastic surface for 16 h. The cells were fixed with fixation and permeabilization reagent (eBioscience, Thermo Fisher Scientific) and incubated either with vinculin (2 μg mouse mAb/ml, clone 7F9, Santa Cruz Biotechnology, Dallas, TX), with paxillin (2 μg mouse mAb/ml, clone B2 Santa Cruz Biotechnology) to reveal FA, with vimentin (3.6 μg mouse mAb/ml, clone 1/9, Dako Products, Agilent, Santa Clara, CA) to label the IFs, or with the MitoTracker™ Red CMX Ros (100 nM, Molecular Probes, Thermo Fisher Scientific) to reveal mitochondria. Chamber slides were incubated with the second ab, a goat-anti-rabbit polyclonal Fab fragment Ab labeled with AF-488 (3 μg/ml, Jackson Laboratories, Bar Harbor, MN) after two times of washing with PBS. For the labeling of f-actin MSCs were counterstained with AF-488 or AF-594 phalloidin (0.1 U/ml, Molecular Probes, Thermo Fisher Scientific) and finally nuclei were stained with DAPI (Sigma-Aldrich). The slides were mounted with Fluoromount-G™ (Southern Biotechnology, Thermo Fisher Scientific) and analyzed with an alpha-Plan-Apochromat 63x objective and a Leica TCS SP8 confocal microscope (Leica Microsystem GmbH, Wetzlar, Germany). Serial dilutions of each primary and secondary antibody were tested to minimize non-specific adsorption, assure separation of the fluorescent signals, and optimize fluorophore concentration to preclude self-quenching.
MSCs were grown, stained and scanned according to the “confocal microscopy” part. The presence of dorsal or ventral SFs as well as transverse arcs was analyzed by using the analysis software StrataQuest (TissueGnostics, Vienna, Austria). Context based quantitative analysis of fluorescence images on an automatic interface segregated dorsal SFs anchored only at one end to FAs, from ventral SFs spanning f-actin filament bundles from two FAs, and from transverse arcs that are not at all anchored to FAs but instead connected to intracellular structures.
All samples were prepared in ibidi 35 mm μ-dishes (Ibidi). First of all, the μ-dishes were coated with 5 μg/cm2 fibronectin (Sigma-Aldrich). The coating was done by covering the surface with a fibronectin solution and incubation for 60 min at 37°C. Thereafter, the remaining solution was discarded. MSCs were put on the fibronectin coated surface and incubated for 24 h at 37°C in 5% CO2 air humidified environment and subsequently immobilized on the surface with 4% formaldehyde (Thermo Fisher Scientific). After immobilization the formaldehyde was discarded, and the μ-dish was filled with 1 ml PBS. A 6000 ILM AFM combination (Keysight, Santa Rosa, CA) was used for all measurements. This system consisted of an inverted microscope (Zeiss Observer D1), a motorized AFM stage (Keysight 6000) and the AFM head itself (N9583A model). The implemented camera was a Hamamatsu C11440. Noise cancellation was done using an Accurion Halcyonics Vario active vibration isolating system combined with an Accurion acoustic damping hood. Calculation of the Young's moduli was done with PicoView 1.18.2 software (Keysight). Further image processing (flattening, statistical evaluation, multilayer images) was performed using Gwyddion 2.45 and PicoImage 5.1.1 software (Keysight). The surface topography of immobilized MSCs was investigated using Bruker MSCT cantilever (Bruker AFM probes, Camarillo, CA). Images were taken in PBS and recorded in contact or tapping mode with 0.5 lines per second. For each sample at least three different cells were measured. The elasticity of immobilized MSCs was investigated by using Bruker MSCT cantilever and by performing force volume measurements in PBS. For most measurements a grid of 32 × 32 points was used. On each measurement point one force distance cycle was performed and the Young's modulus was simultaneously calculated using the Hertz model (Hertz,
MSCs were lysed with 1% protease phosphatase inhibitor in RIPA extraction buffer and, total protein was measured using a Agilent Protein 230 kit (Agilent Technologies). Protein was loaded on precast 4–12% polyacrylamide Bis-Tris gels in a NuPage MOPS-buffer and SDS-PAGE western blotting was performed using a XCell SureLock™ Mini-Cell and a PowerEase 500 W power supply. For the transfer on nitrocellulose membranes with a 0.2 μm pore size a XCell II blot™ module was used (Invitrogen, Thermo Fisher Scientific). Membranes were subsequently blocked with non-fat dry milk (Biorad, Hercules, CA). A primary anti-IDO Ab (0.2 mg, clone H-11, Santa Cruz Biotechnology) was detected with HRP conjugate and Clarity Max Western ECL blotting substrate (Biorad) and the chemiluminescent signal was recorded with a Chemi-Doc documentation system (Biorad) for semi-quantification. For internal control, purified IDO protein (Bio-Techne, R&D Systems, Minneapolis, MN) was used at a total concentration of 25–200 ng.
Human Cytokine profiling was performed with a Proteome Profiler Human XL Cytokine Array Kit (R&D Systems, Minneapolis, MN). MSCBM™- medium (Lonza Group Ltd.) with 8% HPL (MacoPharma) and 3 μl Heparin (5,000 IU/ml, Gilvasan Pharma, Vienna, Austria) was used as sample, the procedure was performed according to the manufacturer's instructions. The chemiluminescent signal was recorded with a Chemi-Doc documentation system (Biorad).
Total RNA was extracted from MSCs cultivated in 2D and 3D with and without addition of 50 ng/ml TNF-α and 50 ng/ml IFN-γ using an RNeasy Mini Kit (Qiagen, Venlo, Netherlands) according to the manufacturer's manual. RNA quality was assessed, and the quantity was measured by an RNA 6000 Nano assay (Agilent Technologies). RNA was adjusted to the same concentration for each sample after measuring the RNA concentration. Then, RNA samples were used for cDNA synthesis with the High-Capacity cDNA Transcription Kit (Applied Biosystems, Thermo Fisher Scientific) in a 20 μl reaction mixture according to the manufacturer's protocol. Primers (Metabion—International AG, Planegg, Germany) for HPRT (Toegel et al.,
Cultivated MSCs were prepared in 8-well Lab-Tek™ chambers (Nunc, Thermo Fisher Scientific) with two different cultivation media. The cytoskeleton buffer with sucrose (CBS) (Symons and Mitchison,
Data were expressed as mean ± standard deviation unless otherwise stated. Mann-Whitney test or unpaired
Three batches of HPL from MacoPharma, used for all experiments, were analyzed with a Proteome Profiler Human XL Cytokine Array with 102 cytokines and results are given in a heat map (
Amnion derived MSCs propagated in adherence to plastic surface and analyzed in passage (P)1 or MSCs from spherical aggregates were positive for MSC specific markers CD73, CD90, and CD105 when cultivated in MSCGM™- medium or MSCBM™ with 8% HPL, and showed a size of ~30–60 μm in diameter (
Effect of HPL on 2D and 3D cultures.
SF formation in response to HPL was investigated by Laser Scanning Microscopy in single 2D adherent P1 MSCs or MSCs that emanated from 3D spherical aggregates (
Mitochondrial distribution in single adherent MSCs showed that mitochondria were found in areas of high cytoplasmic activity and appeared condensed to rods and networks, according to fusion and fission dynamics (
Context-based analysis of confocal images allowed us the quantification of the SFs and the segregation of SFs in ventral SFs, dorsal SFs, and transverse arcs according to their connection with FAs (
Examination of cytoskeleton morphology.
Topographic images, to identify sub-membrane cytoskeletal structures in adherent P1 and P3 MSCs were performed by the AFM in deflection mode. MSCs showed a flat and spread morphology size (
Examination of elasticity.
We used an ImageJ macro tool (Valente et al.,
Mitochondrial pattern analysis and super-resolution imaging.
MSCs were stained with an anti-mitochondria mAb, instead of MitoTracker™, to enable high resolution analysis of ATP5H mitochondrial synthase by a modified Olympus IX81 inverted epifluorescence microscope (
Gene and protein expression of immunological molecules that mediate MSC functions and might rely directly on the status of SF formation and mitochondrial activity like GARP and IDO were analyzed before and after the stimulation with pro-inflammatory cytokines IFN-γ and TNF-α in 2D and 3D cultivated MSCs. Gene as well as IDO protein expression could only be found after stimulation with pro-inflammatory cytokines TNF-α and IFN-γ and not in unstimulated cultures. Due to high variance in IDO RT-PCR the normalization to two housekeeping genes HPRT and PPIA showed no increase in MSCs cultured in MSCBM™ with 8% HPL after 24 and 48 h of stimulation (
Expression of the immunological molecule IDO.
Expression of the immunological molecule GARP.
Extensive research is needed to understand the mechanism how MSCs contribute to organ repair in order to rescue damaged tissue. After application, MSCs can (1) repair connective tissue disease or trauma by integration and differentiation into the required organ specific mature cells at the target site. Furthermore, (2) a locally released secretome of the therapeutic MSCs within the disease milieu can further lead to recovery of damaged organ cells. Finally, (3) a transfer of mitochondria from MSCs to damaged cells in the immediate surroundings by tunneling nanotubes can also facilitate regeneration. The distribution and direct impact of the three different mechanisms of MSC-mediated tissue repair is currently under debate as well as the efficacy of MSC-induced regeneration without post-transplant integration just by the local release of the MSCs' secretome and mitochondrial transfer. Here we investigate the importance of culture medium composition involving HPL as human-derived FBS alternative for MSC expansion. In addition, we investigate the impact of 2D cultivation in adherence to plastic surfaces and 3D cultivation in MSC spherical aggregates generated by hanging drop technology. We found that MSCs grown in adherence to plastic surfaces showed higher levels of f-actin protein and formed thick and well-defined bundles of SFs, while MSCs from spherical aggregates showed lower amounts of f-actin protein which correlated with thinner and more delicate f-actin filaments. The application of HPL led to a reduction in f-actin protein in MSCs grown in adherence to plastic surface while HPL had no effect in MSCs grown in spherical aggregates. It seems likely that HPL has a direct effect on 2D cultured MSC reducing actin stress fiber formation and mitochondrial mass due to the spectrum of cytokines, chemokines, and growth factors provided in the HPL preparation. These human derived bioactive factors were shown to be capable to modulate resiliency of cultured MSCs stressed with palmitate and could lower overall variance in MSC performance between donors (Boland et al.,
Mitochondria are the powerhouses of the cell producing the majority of ATP through respiration and oxidative phosphorylation (OXPHOS) but besides energy generation, mitochondria also participate in calcium signaling, redox homeostasis and apoptosis, thereby mitochondria metabolism can predetermine the fate of transplanted MSCs. It was shown previously that mitochondria accumulation at sites within the cytoplasm of high-energy demand where local ATP production is essential for virtually all cellular functions. Mitochondrial fusion and fission dynamics that involves transportation by IFs is therefore an indicator of MSC fitness. Here we can show that the mitochondrial mass increased during cultivation of adherent MSCs from P1 to P3 as determined by flow cytometry and HPL culture supplementation prevented this increase. Mitochondria were either determined by MitoTracker™, a vital fluorescent lipophilic cationic dye, or by a specific mAb that determines ATP5H mitochondrial synthase and analyzed by 3D reconstructed high resolution immunofluorescence microscopy. We could distinguish between unbranched punctate, rods, and branched structures (networks) with high precision in living MSCs. Unbranched punctate and networks were shown to be increased in adherent P3 MSCs compared to adherent P1 MSCs cultivated in MSCGM™- medium. When P3 adherent MSCs were cultured in HPL, cells showed the same mitochondrial morphology as P1 MSCs. When networks were analyzed for length of branched mitochondrial structures, we found the highest values in P1 adherent MSCs cultured in MSCGM™- medium. When P3 adherent MSCs cultivated in MSCGM™- medium were analyzed by dSTORM and calculated by ImageJ we found a trend toward a decrease in unbranched punctate mitochondria in P3 adherent cultivated in HPL, confirming previous experiments using the MitoTracker™. Interestingly, mitochondrial rods that count for the majority of mitochondrial structures found and mitochondrial network structures that appear as physically interconnected networks showed no dependence on cultivation in HPL when analyzed by dSTORM. dSTORM has the potential to examine mitochondrial reconstruction and semi-automated analysis of immunofluorescent images with high fidelity. It was shown previously that cultivation-induced mitochondrial remodeling involves mitochondrial fission that plays a crucial role in the segregation of aberrant mitochondrial fragments from the remaining tubular network as well as mitophagy. Mitochondria fission is regulated by cytoplasmic proteins Drp1 and Fis1 localized on the outer membrane of mitochondria, while NIX, BNIP3, FUNDC1, PINK (PTEN-induced putative kinase), and PARK2 (E3-ubiquitin ligase Parkin) dependent mechanisms are involved in determining damaged mitochondria and forward them for mitophagy (Bragoszewski et al.,
MSCs facilitate immunomodulatory capabilities involving a network of regulatory pathways. These pathways need to be orchestrated both spatially and chronologically and rely on a dynamic cytoskeleton as well as energy in order to function properly, but the precise mechanism and the regulatory molecules involved are still an area of debate. Here we have selected two prominent immunomodulatory molecules indoleamine-2,3-dioxygenase (IDO), a tryptophan-catabolizing heme-containing enzyme that facilitates a rapid consumption of tryptophan from the local microenvironment and glycoprotein A repetitions predominant (GARP) a membrane receptor binding latency-associated peptide (LAP)/TGF-β1 to study their dependence on cultivation in 2D or 3D and the effect of HPL. IDO is of particular importance because removal of L-tryptophan by IDO starves the environment of this essential amino acid and stress-response pathways such as GCN2 and mTOR signaling pathways are sensitive to amino-acid withdrawal. The biologically active tryptophan catabolites such as kynurenine metabolites can contribute to a
In conclusion, we can say that cultivation in 3D is beneficial for MSCs to maintain low SF formation and physiological membrane elasticity as well as ensure adequate mitochondrial function and ATP supply. Cultivation in 2D, where MSCs adhere to the plastic surface, led to the intensive ventral SF formation where bold f-actin bundles span the entire length of the MSC and perinuclear mitochondrial accumulation. Ventral SFs anchored at two sides to FA can set up an internal force of contraction which enhances membrane elasticity. Serum substitution with HPL could partially reverse this effect in 2D cultivated MSCs leading to lower levels of f-actin protein and mitochondrial mass. For a successful therapeutic application of MSCs in a clinical setting, clinicians need to understand the complex regulatory pathways influencing MSC cyto-architecture, since generation of inner forces induced by ventral SF formation has an effect on membrane elasticity, deformability, and ability of MSCs to migrate. These physical properties of MSCs combined with mitochondrial dynamics can determine the immunomodulatory function of MSCs. Here we can show that IDO was not expressed in unstimulated MSCs and relayed on stimulation with IFN-γ in both 2D and 3D MSC cultures. The immune regulatory molecule GARP was found to be constitutively expressed independent of 2D or 3D cultivation with no additional upregulation upon stimulation with IFN-γ and TNF-α. Interestingly, HPL had no effect on IDO or GARP expression in our 2D or 3D MSC culture systems. This knowledge is relevant because MSCs, upon administration, can migrate to sites of injury, eventually engraft and differentiate into functional mature organ cells, but definitively secrete cytokines, chemokines, hormones and growth factors as well as extracellular vesicles at the target site. To perform combined investigations on SF formation, mitochondrial morphology and deformability of the MSC membrane and link these results with functional properties like the expression of immunomodulatory molecules is important prior to administration and a prerequisite to translate these results in a humanized animal model in the future.
The datasets generated for this study are available on request to the corresponding author.
The studies involving human participants were reviewed and approved by EK791/2008, EK1192/2015, GS1-EK-4/312-2015. The patients/participants provided their written informed consent to participate in this study.
MF conceived the presented idea and supervised the project as principle investigator. MP conducted the experiments and was responsible for acquisition, analysis, and interpretation of data for the work. MP drafted and MF contributed to the final version of the manuscript. VW revised the draft critically for important intellectual content. ER contributed at cell isolation, acquisition, and interpretation of data. CM and AE assisted with the atomic force microscopy and helped carry out the analysis. FH and JJ contributed the 3D dSTORM analysis.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors thank the company TissueGnostics for technical assistance and the Austrian Cluster for Tissue Regeneration for networking.
The Supplementary Material for this article can be found online at: