An integrated bioprocess for the expansion and

Research Article An integrated bioprocess for the expansion and chondrogenic priming of human periosteum derived progenitor cells in suspension bioreactors†

Priyanka Gupta1,2, Liesbet Geris1,3,4, Frank P. Luyten1,2,* & Ioannis Papantoniou1,2,*

1.

Prometheus, Division of Skeletal Tissue Engineering, KU Leuven, O&N1 Herestraat 49, Leuven, Belgium

2.

Skeletal Biology and Engineering Research Center, KU Leuven, O&N1 Herestraat 49, Leuven, Belgium

3.

Biomechanics Research Unit, GIGA-R In Silico Medicine, Université de Liege, Quartier Polytechnique 1, Allée de la découverte 13A, Liège, Belgium

4.

Biomechanics Section, KU Leuven, Celestijnenlaan 300 C (2419), Leuven, Belgium

* These authors share senior authorship Keywords: Autologous human periosteum derived stem cells, Chondrogenic priming, Expansion, Microcarrier, Spinner flask.

†This

article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: [10.1002/biot.201700087].

This article is protected by copyright. All rights reserved Received: June 20, 2017 / Revised: September 1, 2017 / Accepted: October 5, 2017

1

 

Abstract The increasing use of microcarrier based suspension bioreactors for scalable expansion of adult progenitor cells in recent years reveals the necessity of such approaches to address bio manufacturing challenges of advanced therapeutic medicinal products. However, the differentiation of progenitor cells within suspension bioreactors for the production of tissue modules is of equal importance but not well investigated. This study reports on the development of a bioreactor based integrated process for expansion and chondrogenic priming of human periosteum derived stem cells (hPDCs) using Cultispher S microcarriers. Spinner flask based expansion and priming of hPDCs were carried out over 12 days for expansion and 14 days for priming. Characterization of the cells were carried out every 3rd day. Our study showed that hPDCs were able to expand till confluency with fold increase of 3.2±0.64 and to be subsequently primed towards a chondrogenic state within spinner flasks. During expansion, the cells maintained their phenotypic markers, trilineage differentiation capabilities and viability. Upon switching to TGF-β containing media the cells were able to differentiate towards chondrogenic lineage by clustering into mm-sized macrotissues containing hundreds of microcarriers. Chondrogenic priming was further evidenced by the expression of relevant markers at the mRNA level while maintaining their viability. Ectopic implantation of macrotissues highlighted that they were able to sustain their chondrogenic properties for 8 weeks in vivo. The method indicated here, suggests that expansion and relevant priming of progenitor cells can be carried out in an integrated bioprocess using spinner flasks and as such could be potentially extrapolated to other stem and progenitor cell populations.

2

 

1. Introduction The field of Tissue Engineering and Regenerative Medicine (TERM) aims to produce progenitor and stem cell populations as therapeutic substances, rather than being a means to produce therapeutic substances. Due to their relatively easy accessibility and fewer regulatory requirements in comparison to pluripotent stem cells, mesenchymal stem cells (MSCs) are considered to be one of the most important sources of cells for such treatment modalities. On receiving proper cues, MSCs differentiate in vitro into chondrocytes, osteoblasts, adipocytes and even tenocytes, myoblasts and neuronal cells [1-6]. This is further highlighted by the fact that till June 2015 there were around 500 trials registered at ClinicalTrials.gov [7] while the current number, as of May 2017 was more than 700. The current practice for expansion of these cells mainly depends on 2D culture systems like T flasks, multi-layer vessel stacks etc. However for practical use in clinical and pre- clinical studies it is estimated that the number of cells needed can go up to billions. For example, for cartilage regeneration in osteoarthritic cases, around 15 – 45 million MSCs may be required per patient [8, 9], while for long bone defects a 10-fold increase of this amount of cells would be needed for implantation [10] . The use of large scale culture systems or bioreactors are therefore essential in such cases. Chondrogenic differentiation of MSCs is integral for cartilage regeneration as well as for bone fracture healing via endochondral ossification. To date, various culture methods for chondrogenic differentiation have been developed including micromass and pellet systems (both based on cell aggregation), as well as biomaterial-based culture. The use of pellets has been used widely due to the fact that it allows extensive cell-cell interaction and cell-extracellular matrix interactions crucial for chondrogenic differentiation and maturation [11, 12]. However, several groups have reported the presence of undifferentiated and/ or 3

 

necrotic centres in pellet cultures [13, 14]. Micromass cell culture is an important approach for chondrogenesis and has been shown to be a better option in comparison to pellet systems [11, 15]. Biomaterial based systems with or without use of dynamic culture conditions have been used in few cases of differentiation of stem cells into the chondrocytic lineage [16, 17]. However from the translational perspective the aforementioned methods are not easily scalable and are mostly useful in lab-based research following manual procedures. Bioreactors allow scaling up but also provide enhanced nutrient and oxygen supply to the cells, facilitating removal of waste material and maintaining culture homogeneity. Spinner flasks in conjunction with microcarriers are one of the oldest technologies of large scale cell culture. This platform is useful for MSCs because of their anchorage dependence mode of growth. Additionally, the use of microcarriers could serve as an MSC delivery system for tissue engineering [18]. One of the earliest studies reporting the use of microcarriers and spinner flasks for culturing human MSCs was reported by Yu et al., in 2009 [19] wherein human placenta derived MSCs were expanded in stirred reactors using Cytodex3 microcarriers. Since then, a range of MSCs derived from different sources were grown in stirred reactors using microcarriers of different materials [20-28]. Biodegradable collagen based Cultispher S has been used by many groups to simplify this process and hence is widely used for human MSCs culture [22, 27, 29-32] and has also been used in human in vivo as dermal fillers and on venous leg ulcers [33-35]. The majority of studies using spinner flasks for mesenchymal progenitor cell culture has focused on cell expansion aiming to achieve sufficient numbers for clinical practice [26, 28, 36, 37]. Efforts have also been made by some groups to study the effect of using microcarriers and bioreactors for differentiation of MSCs in to different lineages primarily towards the osteogenic lineage [16, 32, 38, 39]. 4

 

However, very few groups have tried to expand and subsequently differentiate and/or prime different stem cells within the same system using specific media and reactor setups. Human embryonic stem cells were expanded and differentiated towards endodermal lineage [40] while mouse induced pluripotent stem cells were expanded and differentiated towards cardiomyocytes [41] in spinner flask based cultures. Human embryonic stem cells were also expanded and differentiated towards cardiomyocytes and hepatic lineage [42, 43]. Chondrogenic differentiation of spinner flask expanded MSCs was carried out by modulating the cell cytoskeleton within a dynamic system [16]. That publication also highlighted the importance of selecting correct attachment surfaces for proper differentiation of cells. Within in vivo bone fracture healing models, periosteal derived cells have been shown to assist enormouslybe the main cell source in long bone fracture healing via endochondral ossification participating in the in vivo formation of the callus, a fibrocartilaginous provisional tissue, that is subsequently mineralised forming and replace de novo by bone[44, 45]. In contrast bone marrow MSCs (BM-MSCs) are known to assist only to a limited extent in long bone fracture healing via intramembranous ossification [44, 46, 47]. In a comparative study of MSCs from different sources, it has also been reported that PDCs have a higher proliferative capacity in comparison to BM- MSCs as well as the second highest in vitro chondrogenic potential next to cells from synovial membrane [46, 48]. Periosteum derived stem cells (PDCs) have been known to be important progenitor cells for treatment of skeletal defects and injuries while after ex vivo expansion, they have been shown to have characteristics similar to mesenchymal stem cells derived from other sources [49]. In addition they have shown to possess superior bone forming properties to bone marrow derived and other MSCs when implanted on specific Calcium Phosphate (CaP) carriers [46]. When cultured in perfusion 5

 

bioreactors they have been shown to be able to differentiate into the osteogenic lineage producing mineralized extra cellular matrix (ECM) [50, 51] and to heal calvarial defects in mice [52] . Previously, our group carried out expansion of these cells in large scale bioreactor systems [10, 53], but till now, to the best of our knowledge, there has been no attempt to expand and prime these cells towards chondrogenic lineage in the same culture setup. This study investigated the expansion and priming of hPDCs towards the chondrogenic lineage in a single bioprocess simply by switching the media once confluency was attained (fig.1). 2. Materials and Methods 2.1 Human periosteum derived cells (hPDC) static culture: Human periosteum derived cells (hPDCs) were obtained from periosteal biopsies with informed consent from the patient and approved by the Ethics Committee for Human Medical Research (KU Leuven). Pooled hPDCs from three individual donors were expanded and maintained in high glucose containing Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen) supplemented with 10% Fetal Bovine Serum (FBS, Hyclone), 1% sodium pyruvate (Invitrogen) and 1% antibiotic- antimicotic (Invitrogen). Cells were incubated in a humidified incubator at 37°C with 5% CO2. They were monitored regularly and on reaching a confluency of 90%, they were detached with Tryple E (Invitrogen) and used for further experiments. 2.2 Spinner Flask Culture: 100 ml spinner flasks with paddle based impeller from Bellco Biotechnology (Bellco Glass, Inc) were used with a final media volume of 80 ml for expansion phase and 40ml for the differentiation phase. The spinner flasks were coated with SigmaCote (Sigma Aldrich) prior to usage. Cultipsher S microcarriers (GE Health 6

 

Care) at a concentration of 1 mg/ml were weighed, hydrated using PBS and sterilized by autoclaving as per manufacturer’s instructions. The microcarriers were equilibrated prior to use by incubating them in culture medium overnight. Preliminary studies were carried out to test different seeding densities (data not shown) and a seeding density of 2.5x 104 cells/ml was selected. Three different attachment protocols were initially tested (data not shown) – (a) continuous agitation at 30RPM for 24 hrs, (b) intermittent agitation protocol of 5 mins ON at 30 RPM and 2 hrs OFF for 7-8 hrs followed by overnight agitation at 30 RPM and (c) 8 hrs continuous agitation at 30 RPM followed by overnight static culture. Selection of attachment protocol was based on seeding efficiency. The 24hr intermittent agitation protocol showed the highest efficiency and was adopted, wherein the cells and microcarriers were inoculated in 50% of the total medium volume (40 ml) in spinner flasks and an intermittent agitation protocol was followed for 7-8 hrs. The flasks were then kept overnight at a constant spin rate of 30 RPM. After 24 hrs, the volume was increased to 100% (80ml) and the spin rate to 50 RPM. The hPDCs were expanded for 12 days in the spinner flask with 50 % medium replacement every 3rd day. On the 12th day, the expansion medium was replaced by chondrogenic differentiation medium described later, the volume was reduced to 40ml and the cells were then cultured for another 2 weeks. At specific time points throughout the experiment, cells and media were stored for different analyses. 2.3 Cell count and DNA quantification: DNA quantification and viable cell count were carried out for estimation of cell expansion rate and viability. Cultispher S microcariers were dissolved using collagenase Type IV (Life Technologies) in order to harvest the cells. Trypan Blue exclusion assay was used to count viable cells. For DNA quantification, cells were 7

 

collected in lysis buffer (Qiagen) supplemented with β mercaptoethanol and vortexed for 30 secs. DNA quantification was carried out using Qubit ds DNA HS assay kit (Molecular Probes, Thermo Scientific) following manufacturer’s protocol using Qubit Fluorometer (Invitrogen). 2.4 Medium Analysis: Metabolites like glucose, lactate, ammonia, and lactate dehydrogenase were measured using Cedex Bio Analyzer (Roche). Collected medium was centrifuged at 3000 rpm for 10 mins in order to remove cell debris. The supernatant was then collected and analysed. Ammonia concentration was normalised with respect to its production in cell free set up due to glutamine degradation prior to analysis. Specific metabolite consumption/production rate was for expansion and priming phase using equations 1 and 2 respectively. 𝑞𝑞𝑚𝑚𝑚𝑚𝑚𝑚 =

𝑞𝑞𝑚𝑚𝑚𝑚𝑚𝑚 =

𝐶𝐶𝑚𝑚𝑚𝑚𝑚𝑚(𝑡𝑡) −𝐶𝐶𝑚𝑚𝑚𝑚𝑚𝑚(𝑡𝑡−1) 𝑋𝑋×𝑡𝑡

(1)

𝐶𝐶𝑚𝑚𝑚𝑚𝑚𝑚(𝑡𝑡) −𝐶𝐶𝑚𝑚𝑚𝑚𝑚𝑚(𝑡𝑡−1) 𝑡𝑡

(2)

Where, 𝑞𝑞𝑚𝑚𝑚𝑚𝑚𝑚 = cell specific metabolite consumption/production rate,

𝐶𝐶𝑚𝑚𝑚𝑚𝑚𝑚(𝑡𝑡) = concentration of metabolite measured at the end of time point t,

𝐶𝐶𝑚𝑚𝑚𝑚𝑚𝑚(𝑡𝑡−1) = Concentration of metabolite at the start,

𝑋𝑋 = Cell number at time t, t = time (day)

∆t = time (t-1) - time(t)

2.5 Live Dead Assay:

8

 

Analysis of live and dead cells was carried out using Live/Dead Viability/Cytotoxicity kit (Molecular Probes, Thermo Scientific). Briefly, cell samples were taken from the spinner flask and washed twice with PBS. 0.5µl of Calcein –AM (4mM stock) and 2µl of Ethidium Homodimer (2mM stock) was added to 1ml of PBS. About 100-300µl of the solution was added to the samples (depending on their size) and incubated at 37°C for 20 – 30 mins. The solution was then removed and the samples were washed in PBS twice. They were then observed and photographed using a fluorescence microscope (Discovery V8, Zeiss). 2.6 DAPI, Phalloidin staining: At specific time points, cells on microcarriers were collected from the spinner flask for actin filament and nucleus staining. The cells on microcarrier were fixed using 4% PFA and washed with PBS twice. They were then incubated with 0.1M Glycine solution for 15 mins to reduce background staining followed by PBS wash again. They were incubated for 20 mins at room temperature in staining solution containing 2.5µl of DAPI (1mg/ml stock solution), 4µl of Phalloidin (200 U/ml stock concentration), Alexa Fluor 488, Phalloidin, Life Technologies) and 20µl of Triton X per ml of PBS. Cells on microcarriers were then imaged using confocal microscope (LSM 780 – NLO Confocal Microscope, Zeiss). 2.7 RNA extraction, cDNA synthesis and qPCR analysis: Total RNA from specific volume of cells was extracted using RNeasy mini kit (Qiagen) as per manufacturer’s instructions and concentration was quantified using NanoDrop ND-1000 spectrophotometer (Thermo Scientifics). cDNA synthesis was carried out using RevertAid H Minus 1st strand cDNA synthesis kit (Thermo Scientifics) for 500ng of RNA. cDNA was stored at - 20°C at a final concentration of 3.33ng/µl. qPCR was carried out using SyBR Green Mastermix 9

 

(Thermofisher Scientifics). The PCR cycle used was - 45°C for 2 mins, 95°C for 30 secs 40 cycles of 95°C for 3 s, and 60°C for 20 s. Each sample was tested in duplicate and compared with β-actin expression that allowed normalization of results. Relative differences in expression were calculated using the 2-ΔΔCT method. 2.8 MSC Phenotyping: Harvested cells were assessed based on a combined positivity for typical MSC cluster of differentiation (CD) markers and lack of expression for hematopoietic markers. The used antibody panel was CD90-FITC, CD73-APC, CD105-PE, CD14, CD20, CD34, CD45-PerCP (Miltenyi Biotec). Dead cells were excluded based on a viability dye. Flow cytometric analysis was carried out using BD FACS Canto. Antibody titration was performed according to the protocol of Hulspas [54]. Automatic single-color compensation was performed by the acquisition software (BD FACSDiva) using compensation beads (UltraComp eBeads Affymetrix eBioscience), except for the viability dye as it works based on internalization of the dye in the cell. The gating was based on FMO (Fluorescence Minus One) controls [55], except for the viability dye where the gating was based on the signal of a 50-50 mixture of healthy cells and heat/cold-killed cells. 2.9 Trilineage in vitro differentiation potential analysis: In order to assess the multipotency maintenance capability of the cells, their trilineage differentiation capability in vitro was tested. Chondrogenic differentiation of the hPDCs was assessed in a micromass assay as described earlier [49]. Briefly, 2 × 105 cells were re-suspended in 10µl of regular culture medium and seeded as micromasses in a 24-well plate. After 2 hrs of incubation, 0.5 mL of standard culture medium was added. After overnight 10

 

incubation, the medium was replaced by chondrogenic medium consisting of DMEM/F12 (Life Technologies), 2% FBS, 1% antibiotic–antimycotic, 1% ITS Premix (Corning), 100 nM dexamethasone (Sigma), 10 M Y27632 (Axon Medchem), 50 g/mL ascorbic acid, 40 g/mL proline and 10 ng/mL recombinant human transforming growth factor β-1 (Preprotech). The chondrogenic medium was refreshed every other day. Micromasses in normal culture medium were taken along as negative control. Preliminary comparative study between media containing only TGF β and our chondrogenic media showed that only TGF β was unable to differentiate hPDCs towards chondorgenic lineage. After 7 days of chondrogenic induction the micro-masses were fixed in ice cold methanol and stained at room temperature for 1 hr with a 0.1% alcian Blue solution in 0.1 M HCL at pH 1.2. Alcian stain was extracted using 6M guanidine hydrochloride and absorbance was measured at 620nm. Osteogenic differentiation was assessed by seeding the cells in 24-well plates at a density of 4500 cells/cm2 in 0.5 mL standard culture medium. After 48hrs the medium was replaced by standard culture medium supplemented with 100 mM dexamethasone, 50 g/mL ascorbic acid and 10 mM β-glycerolphosphate. The medium was refreshed every 2 days for 21 days. Samples in standard culture medium were used as negative control. Cells were fixed prior to analysis in ice cold methanol for 1 h and afterwards stained with a 2% Alizarin Red S solution in Baxter water. 10% cetylpyridinium chloride was used for stain extraction and measured at 540nm. Adipogenic differentiation was investigated by seeding the cultured hPDCs in 24well plates at a density of 1 × 104 cells/cm2 in 0.5 mL standard culture medium. After 24 hrs, the medium was replaced by adipogenic medium consisting of α MEM (Life Technologies) supplemented with 10% FBS, 1% antibiotic– antimycotic, 1 M dexamethasone, 10 g/mL human insulin (Sigma), 100 M 11

 

indomethacin (Sigma) and 25 M 3-isobutyl-1-methylxanthine (Sigma). Medium was refreshed 2 days till visible fat globules are seen. Cells were then fixed in 10% formaldehyde for 20 min, rinsed shortly with 60% isopropanol and stained with Oil Red O. Staining from fat globules was extracted using isopropanol and measured at 518nm. 2.10 Ectopic in vivo implantation and analysis: Selected macrotissues from spinner flasks after priming were ectopically implanted in nude mice for 8 weeks to validate the propensity of the cells to maintain their chondrogenic priming in vivo. Only microcarriers were encapsulated in collagen gel and implanted as control. After 8 weeks, the explants were fixed, embedded, sectioned and analysed for Glycosaminoglycan (GAG) content by alcian blue staining. 2.11 Statistical Analysis: All quantitative results were expressed as mean±SEM. One Way Anova followed by Dunnet's multiple comparison test was performed using Graph Pad Prism, version 6.00 for Windows (GraphPad Software) in order to find statistical significance of the data. Statistical significance analysis was carried out by comparing the experimental setups with day 0 (D0) control set-ups. Exceptions are mentioned within the results. Differences among data were considered to be statistically significant if p