Community Research and Development Information Service - CORDIS

FP7

PURSTEM Report Summary

Project ID: 223298
Funded under: FP7-HEALTH
Country: Ireland

Final Report Summary - PURSTEM (Utilisation of the mesenchymal stem cell receptome for rational development of uniform, serum-free culture conditions and tools for cell characterization)

Executive Summary:
Stem cells offer a promising avenue to therapy for a wide range of human diseases, such as osteoarthritis. However, for this potential to be realized, a consistent and plentiful supply of well-characterised stem cells is essential. To date, there has been relatively little progress in the development of new culture technologies for the large-scale manufacture of mesenchymal stem cells (MSCs). Current obstacles to the large-scale manufacture of MSCs suitable for therapeutic use in humans include:
• a lack of standards for the characterisation, isolation or identification of MSCs
• the absence of standard protocols for differentiation of MSCs to various lineages
• a lack of specificity for surface markers used for MSC characterisation
• the absence of standardised cryopreservation protocols
• the requirement of current production methods to use animal-derived serum.

The PurStem project set out to identify the MSC “receptome” in order to develop novel serum-free media suitable for large-scale MSC production. Furthermore, the PurStem project aimed to create novel antibody reagents for specific MSC characterization and to contribute to Good Manufacturing Practice (GMP) standards to enable rapid progression to production of serum-free MSC for clinical, therapeutic applications.

The PurStem project has achieved its goals by:
• developing and validating a collaborative, standardised procedure for the isolation, culture and cryopreservation of MSCs. These procedures produced consistent cell types even when grown in different laboratories
• using the wealth of information obtained from characterisation of the MSC “receptome” to develop low-serum and serum-free culture conditions where MSCs can survive and proliferate, thereby reducing or eliminating potential contamination issues associated with current serum-based cell culture methods
• using a novel combination of transcriptional and biochemical approaches to identify specific surface proteins that can isolate MSCs from other cell types
• identifying new and improved reagents and methods for directly isolating MSCs from tissue based on immunoselection using antibodies and producing germ-free antibodies for GMP-grade isolation of human MSCs for clinical applications.

These results will contribute to the optimisation of GMP manufacturing and banking of cells, initially for use in clinical trials and ultimately as a commercial product. Furthermore, PurStem has also advanced our understanding of MSC, setting the stage for the development of next generation therapies to exploit the self-repair potential of adult stem cells or stem cell targeting.

Stem cell therapeutics are expected to lead to the treatment of diseases with no current effective treatment options, as well as contributing to tissue engineering for replacement purposes. The findings of PurStem will support the development of promising stem cell therapies, initially in the area of osteoarthritis. These novel regenerative therapies can reduce the economic and social costs of disease to the European Community. The primary societal benefits from PurStem will accrue from improvements in treatment options for a range of diseases. These improved treatments will improve quality of life for patients and reduce the economic burden associated with chronic disease.

Project Context and Objectives:
Stem cells offer a promising avenue to therapy for a wide range of human diseases. However, for this potential to be realized, a consistent and plentiful supply of well-characterised stem cells is essential. To date, there has been relatively little progress in the development of new culture technologies for the large-scale manufacture of mesenchymal stem cells (MSCs). The current state-of-the-art has several weaknesses, in that, there are no standards for the characterisation, isolation or identification of MSCs from any tissue, nor are there standard protocols for differentiation of MSCs to various lineages. Additionally, surface markers used for MSC characterisation lack specificity and cryopreservation protocols are not standardised. Critically, current production methods for MSC require the use of animal products with major contaminant implications.

Therefore, in an effort to overcome these issues, PurStem seeks to identify the MSC “receptome” and use this repertoire of growth factor receptors to develop novel serum-free media for MSC production. PurStem aimed to create novel antibody reagents for specific MSC characterization and contribute to GMP manufacturing standards, which will enable rapid progression to production of serum-free MSC for clinical applications.

The objectives of PurStem were to:
? Examine existing methods and approaches to the preparation of MSCs
? Identify current best practice
? Standardise the technology by means of a unified operating protocol.
? Develop new methods of culture and new media formulations by identifying the repertoire of growth factor receptors that exists on the surface of the cell.
? Use this information to develop new media using a combinatorial approach to the selection of growth factor supplements to modulate and optimise the growth kinetics and differentiation of the cells.
? Utilise recombinant human growth factors and so benefit from freedom of reliance on serum.
? Identify new reagents from this effort that will be used to characterise the cells.

As a first step to achieving these goals, all partners agreed on current best practice MSC methods. Consequently, a collaborative standardised procedure for the isolation, culture and cryopreservation of MSCs was developed and validated in each partner laboratory. Using FACS analysis, preliminary work on identifying the “in vivo” MSC transcriptome and changes due to in vitro culture commenced. In parallel, the repertoire of growth factors expressed on the surface of the MSC was identified. Furthermore, a list of novel antibodies for new surface antigens was characterised using a novel combination of transcriptional and biochemical approaches. With respect to the serum-free media formulation, preliminary studies examined MSC viability in low serum conditions containing a number of growth factors and serum free conditions using growth factors in combination with various attachment factors. In order to advance the bioprocessing of MSCs for clinical applications, initial in vivo studies were conducted using the MSCs isolated and cultured using the PurStem unified protocols. Work was also carried out in a “germ-free” chicken facility for the manufacture of the antibodies. Furthermore, a list of the GMP compliant raw materials was identified and an SOP for the small-scale production of serum free media was written.

A number of key results were achieved over the course of the project. Using the unified protocols for the isolation and characterisation of MSC, initial results revealed that there was variability between partners at the start, but these variations reduced over time. Cumulative population doublings throughout the culture process from primary to the end of passage 2 culture indicated that the culture process was relatively reproducible between partner institutions. Moreover, using FACS analysis, it was revealed that the standardised PurStem culture conditions produced cells with a consistent cell surface phenotype even when grown in different laboratories. Using FACS and real time PCR, the repertoire of growth factor receptors and active transcription factors in the “in vivo” MSC was characterised; initial results revealed that cultured MSCs were a heterogeneous population of undifferentiated and more-committed, less proliferative cells, which evolve during passaging towards cellular senescence.

A wealth of information was generated by the surface receptome analysis, where a list of growth factor receptors expressed on the surface of the MSC was identified. The receptome was used to develop the serum free media. In a further effort to optimise the serum-free media formulation, a range of growth factor combinations were examined, and it was found that MSCs can attach, survive and proliferate in low serum and serum free conditions.

Using a novel combination of transcriptional and biochemical approaches to identify much needed antigens against MSC-enriched surface proteins, approximately 20 proteins were identified that are enriched in the MSC receptome and distinguish MSCs from other cell types within the human MSC niche.
In the end, a rigorous protocol for the isolation and culture of MSCs has been developed and validated. The repertoire of identified growth factors was used to:
• Narrow down the growth factor combinations for the serum free media formulation,
• Provide information on new reagents that can strengthen the release criteria for MSCs,
• Help identify new methods for directly isolating MSCs from tissue based on immunoselection rather than adherence
• Help enable new methods for endogenous manipulation of MSCs.

Furthermore, once specificity of the proteins was confirmed, positive antigen ‘hits' were used for the production of germ-free antibodies for use in the GMP-grade isolation of human MSC for clinical applications. The ability to produce GMP-grade human MSC has advanced the state of the art in MSC standards of preparation and release criteria.

PurStem has generated new standard operating procedures for the isolation and growth of MSCs, as well as new criteria that will be used to define the MSC for current applications in tissue engineering. These will contribute to the optimisation of GMP manufacturing and banking of cells for use first in clinical trials and subsequently as the basis for commercial products. PurStem will contribute to bridging the anticipated gap between demand and supply of consistent cells, as more stem cell therapies are approved by regulatory authorities for use in clinical medicine.

PurStem has contributed not only to optimal stem cell manufacturing processes for Regenerative Medicine applications in the short-term, but has also advanced our basic understanding of MSC biology by defining the surface receptome. This database will improve our knowledge of the pathways that control mobilization and growth of the “in vivo” stem cell in response to injury, setting the stage for the development of next generation therapies which will exploit the self-repair potential of adult stem cells or stem cell targeting.

The area of stem cell therapeutics is expected to lead to treatment of diseases with no current effective treatment options as well as contribute to tissue engineering of new tissues or organs for replacement purposes. The particular interest of PurStem consortium members is in arthritis and bone repair. Osteoarthritis is a disease associated with ageing and is the most common of the rheumatic diseases. PurStem will support the translation of promising stem cell therapies in this area to treatment of the joint injuries that contribute to development of osteoarthritis as well as potential therapies that prevent progression of disease. These novel regenerative therapies will result in reducing the economic and social costs of the disease to the European Community. The primary social benefits from PurStem will accrue from improvements in treatment options in these and other areas. These will drive improved quality of life for patients, reduction in the economic burden associated with chronic disease, reproducible treatments, and a decreased risk of results dispersion and uncertainties.

Since the onset of the PurStem project, a greater awareness of the research required to advance the large-scale manufacture of stem cells for clinical applications is becoming more apparent. In order to highlight the research currently underway in Europe, the PurStem website was created and has been constantly updated over the course of the project. Furthermore, PurStem has made an imprint on the world stage, with several conference presentations in the United States, United Kingdom, Ireland and the Czech Republic.

Project Results:
Standardisation of MSC isolation, expansion, characterisation and banking (WP1)

Why standardise MSC isolation, expansion, characterisation and banking?

Adult human bone marrow derived mesenchymal stem cells (MSCs) hold great promise as clinical treatments for several pathologies due to their multipotentiality. However, given the low percentage of MSCs present in adult bone marrow (approximately 1 in 100,000) compounded by the need for significant numbers of MSCs for clinical treatments, in vitro culturing of MSCs is necessary for clinical application. Currently there is no standardised approach for culturing adult human MSCs. Various combinations of basal medium, supplements and culture conditions can affect the phenotype of the cultured MSC. Each laboratory has its own procedures for MSC culturing, making it difficult to compare published data between laboratories. There was therefore an obvious need for standardised best practice culture methods to be generated between PurStem laboratories, but eventually with translation to the field in general. For these reasons, one of the goals of PurStem was to develop and perform inter-laboratory validation of standardised MSC isolation, expansion, characterisation and banking procedures to achieve the objectives of the PurStem project and to standardise the production of MSCs for clinical use.

PurStem Results

Collaborative Standardised Procedures

Initially, all partners collaborated to develop standardised PurStem procedures for the isolation and culture of MSCs. To ensure that the standardised PurStem procedures were effective and gave similar results for all laboratories, all partners then conducted a laboratory validation. For the validation, bone marrow aspirate was shipped from NUIG to all partners. All partners then followed the standardised PurStem procedures for MSC isolation and culture. Initial efforts to standardise intra-laboratory procedures resulted in variable results, which were addressed by determining optimal storage conditions for shipping and improving the cell counting procedure. A cryopreservation study was also carried out and a GMP-compliant procedure generating cells with a shelf-life of 30 days with cell viability over the 70% GMP threshold was achieved.

Comparison of MSC culturing/differentiation between partner laboratories
Experiments looking at the culture process showed that it was relatively reproducible for all partner laboratories when using the standardised PurStem procedures. The amount of time for the number of cells to double did vary between partners and for different marrow samples. This variability could be explained to a certain extent by the relative number of cells prior to sub-culturing or by an overestimation of the number of MNCs at the start of culturing.

In order for isolated and cultured MSCs to be effective as therapies, they must retain their ability to differentiate into different cell types. Therefore, PurStem partners assessed the ability of the MSCs obtained using the standardised PurStem procedures to form new bone material (osteogenesis), adipose tissue (adipogenesis) and cartilage (chondrogenesis). MSCs from all partners underwent osteogenesis, adipogenesis and chondrogenesis and there was no significant difference between partners.

Comparison of MSC phenotype between partner laboratories
A consistent phenotype (physical appearance and biochemical characteristics) is also important when isolating and culturing MSCs for therapeutic applications. Therefore, PurStem partners examined the phenotypes of MSCs isolated and cultured by each of the partner laboratories using the standardised PurStem procedures. This examination was carried out using fluorescence-activated cell sorting (FACS) analysis to measure markers expressed on the surface of the cells. The ensemble of surface markers expressed by MSCs is termed the MSC “Receptome”.

Pooling the data from each donor at all laboratories showed little variation in the highly expressed markers such as CD73, CD90, CD29. The strongest and most consistently expressed cell surface marker on the cultured cells from all of the partner laboratories was CD73, with all cells being > 90% positive compared to the isotype control. In agreement with the ISCT criteria and previously published data, the cultured cells from all partner laboratories were negative for the expression of CD19, CD45, CD34 and CD31. There is a degree of variation in the expression of CD105, CD146, CD13, CD106, with the most variable markers being CD105 and CD13. This is particularly relevant as the current “standard” of MSC phenotype, the position statement states that MSC must be positive for the expression of CD73, CD90 and CD105. Whilst a proportion of the cells cultured under standardised PurStem conditions did express CD105, the variable expression levels observed between laboratories in these standardised conditions indicates that this is not the most suitable marker for defining MSC phenotype and may in part contribute to the disparity of MSC phenotypes reported in the literature.

Our data indicates that the variation between the different donors is greater than the variation observed between the cells from the same donor being grown in different laboratories. This demonstrates the robustness of our standardised culture technique. The strongest and most consistently expressed cell surface marker on the cultured cells from all donors was CD73.

The marker with the largest amount of variability in expression between different donors was CD13. Expression levels of CD146 and CD105 were also found to be very variable between donors. Variations may be a reflection of the ‘in vitro’ age of the cultures, suggesting that cultures should be compared in terms of the number of population doublings (PD) rather than just passage number.

CD73, CD90, CD166 and CD81 showed no change in expression or variability, being consistently highly expressed at both passages tested. There was no change in the expression or variability for CD106, CD105, CD146 and CD13 at different passages, although all of these markers were highly variable between different donors. There was a trend for expression of CD271 to decrease with passage during culture, although the high donor variability means this was not statistically significant. Based on the two passages tested for cells cultured from the 3 donors, there were no statistically significant differences in the cell surface expression or variability observed between cells of different passages for any of the markers tested.

Effect of extended cultivation on the MSC “Receptome”

Culturing can result in changes in MSCs that may impact their effectiveness for clinical therapies and can lead to a state where cells are no longer able to divide (senescence). Therefore, the MSC “Receptome” was characterised during long-term cultivation under standardised PurStem procedures. A statistically significant negative correlation was observed between the expression of Frizzled 9 (CD349) and TGFbR2 and the number of population doublings. SSEA4 expression showed a similar inverse but non-significant trend. Frizzled 4 (CD344) and TGFbR3 expression was consistently low. FGF receptors were expressed stably, but at relatively low levels. No significant trend was observed for CD106. These data suggest that the expression of Frizzled 9 may serve as the best potency indicator for MSCs cultured using standardised PurStem procedures.

The inter-laboratory variability of “Receptome” marker expression in cells was also studied using cells of varying passages cultured from 4 donors using standardised PurStem procedures. In spite of variation between donors, there was a overall trend of loss of Frizzled 9 expression during culture. There was variation in expression levels between different individuals but also between partners from the same sample. Therefore, Frizzled 9 expression may also be a useful measure of ‘culture quality’.

Additional receptors were examined to assess the degree of variation observed in marker expression between individual donors and between cells from the same donor cultured by different partners. Due to sample constraints, statistical analysis was not possible. However, following trends were observed. For some markers, such as FGFR1, there was a high degree of variation between individual donors and for cells from the same donor cultured in different laboratories. The expression of other markers, including SSEA-4, CD106 and CD146, was highly dependent upon the culturing partner. This was surprising as the goal of standardising the isolation and culture procedures was to avoid culture-related variations. It is possible that seemingly minor differences, such as small variations in culture oxygen levels, may have a greater than expected influence.

1.3.2 In vivo MSC (WP2)

Why characterise in vivo MSCs?
Cultured and in vivo MSCs are heterogeneous populations containing mesenchymal progenitors with differing abilities to form clones (clonogenicity) and differentiate into a number of cell types (multipotentiality). The main difference between cultured and in vivo MSCs is their overall proliferative capacity. Many published studies have also indicated that long-term culture of MSCs is associated with a gradual loss of multipotentiality, including a decline in osteogenesis but this remains controversial. Therefore, some investigators, including PurStem partners UNIGE and UL, have reasoned that in vivo MSCs are likely to be superior to cultured MSCs in both their proliferative and osteogenic capacities. Current inconsistencies in clinical and experimental outcomes of MSC-based therapies may therefore result from a lack of standards to define the in vitro age of the cultured MSCs. Thus, a comprehensive knowledge of in vivo MSC biology is needed in order to fully understand the changes that occur during MSC cultivation and their impact on MSC potency. For these reasons, one of the goals of PurStem was to characterise the repertoire of growth factors and active transcription factors of in vivo MSCs for use as a point of reference against which cultured MSCs could be assessed.

How are in vivo MSCs isolated?

PurStem researchers at UL have pioneered and validated several methods for isolating in vivo MSCs from human bone marrow. Briefly, microbeads are used to first select a large population of cells from bone marrow samples. These cells are then further sorted using flow cytometry combined with a rare-event analysis, yielding in vivo MSCs, known as CD45lowCD271+. Other methods for MSC isolation from bone marrow exist but the methods developed by UL provide better selectivity and higher yields. The UL approach for in vivo MSC isolation has been validated by several independent groups worldwide and is promoted by Miltenyi Biotec, the market leader in cell purification, for which UL serves as a reference laboratory.

PurStem Results

Fluorescence-activated cell sorting (FACS) of in vivo and in vitro MSCs

UL performed a comprehensive analysis of historical and new data comparing MSC marker expression on in vivo MSCs with early-passage cultured MSCs. No differences were found for the expression of the two most commonly used MSC markers, CD73 and CD105. CD166 and CD271, on the other hand, demonstrated significant differences in expression between in vivo and early-passage cultured MSCs. Therefore, CD271 and CD166 markers appear to be specific to in vivo MSCs and are unlikely to have significant value for defining cultured MSCs. CD271 was significantly down-regulated, whereas CD166 was significantly up-regulated. CD146 and CD106 demonstrated the highest donor-to-donor variation in their expression patterns, for both in vivo and cultured MSCs. On average, cultured MSCs had more CD146-positive and CD106-positive cells, however these differences were not statistically significant.

Using samples from a single bone marrow donor, the expression of CD73 and CD105, as well as the proliferative and osteogenic capacities, were investigated using both in vivo MSCs and donor-matched cultured MSCs at 9, 13, 20 and 25 population doublings. Continuous in vitro MSC culture up to 25 population doublings led to a gradual loss in proliferative and osteogenic capacities, however no major differences were observed between in vivo and early-passage MSCs. No differences were found for the expression of standard MSC markers CD73 and CD105. These results confirm that CD73 and CD105 are robust markers for both in vivo and cultured MSCs but that their expression is unrelated to proliferative and osteogenic capacities.

Using samples from three other bone marrow donors, the expression of surface markers after long-term culture (~20-25 population doublings) was studied. Cultured MSCs were uniformly negative for CD45, CD31, CD322/JAM2 and CD271. The expression of CD105, CD73 was generally maintained regardless the culture’s “in vitro age”. Similarly, the expression of two common fibroblast markers (CD90 and CD81) and CD166 was stable. In contrast, both CD146 and CD106 declined in parallel with accrued population doublings.

In vivo MSC transcriptome

UL has previously performed a molecular comparison between in vivo and cultured MSCs using Affymetrix® microarrays that measure particular genes in a biological sample. These microarray experiments were not successful on low cell number samples. Therefore, novel cell lysis/RNA extraction methods were developed during PurStem in order to analyse samples with only 1,000-5,000 sorted cells. A commercial kit from Norgen was used as it could be used to simultaneously obtain RNA, DNA (for telomere work) and protein (to archive for subsequent protein validation work).

Genes related to several differentiation pathways for MSCs (osteo-, adipo-, chondro-, myo- and stroma-genetic), including both early and late differentiation markers, with an emphasis on MSC osteogenic differentiation were selected from the preliminary UL Affymetrix® data for incorporation in a 96-gene microfluidic card. Applied Biosystems microfluidic cards were used due to their effective use of small quantities of biological material. In addition, early-immature markers were also included, to control for undifferentiated MSCs. In vivo MSCs were isolated in parallel with donor-matched hematopoietic lineage cells (HLCs) for use as control cells. Genetic material from the cells was then analysed using the microfluidic card.

Transcription of adipogenesis-related genes was MSC-specific, confirming that at the molecular level, BM CD45lowCD271+ cells were capable of adipogenic differentiation. With the exception of CEBPA, which is known to be the least MSC-specific, all other adipogenesis-related genes demonstrated higher-level expression in in vivo MSCs than HLCs.

Similarly, the overwhelming majority of osteogenesis-related genes were also MSC-specific, confirming that, at the molecular level, in vivo MSCs were capable of osteogenic differentiation. Interestingly, differences in transcription factor RUNX2 expression were not as prominent as that of mature osteogenic markers like BGLAP (osteocalcin) or SPARC (osteonectin). We also identified novel cell surface molecules related to osteogenesis in in vivo MSCs based on several pathways implicated in osteogenesis (TGFb/BMP, Wnt, FGF and IGF pathways). FGFR1 and TGFBR3 were MSC specific, confirming the “osteogenic readiness” of in vivo MSCs.

In vivo MSCs expressed the chondrogenic master-regulator Sox9 but they did not express mature chondrogenic molecules, such as COL2A1 (Collagen 2). Therefore, in vivo MSCs were “ready” for chondrogenesis, but chondrogenic differentiation was not taking place under steady-state conditions in vivo. In contrast, stroma-related genes (CXCL12 (SDF-1), IL7 and others) demonstrated clear MSC-specificity, confirming that at the molecular level, in vivo MSCs were acting as stromal cells supporting hematopoiesis. In fact, all 8 stroma-related genes present on the card showed clear specificity for in vivo MSCs over matched HLCs.
Analysis of molecules associated with very immature stem cells/ES cells showed that with the exception of NES (nestin), other immature markers were non-selective in CD45lowCD271+ cells compared to donor-matched HLCs.

Comparison of transcriptome of in vivo and early passage MSCs

Using the same set of genes as above, the stability of the MSC transcriptome during early passaging under standardised PurStem conditions was examined by comparing the transcriptome of in vivo MSCs with donor-matched early-passage MSCs. No statistically significant differences between freshly isolated in vivo MSCs and early passage MSCs were observed although trends were observed for some markers. Stroma- and adipose-related transcripts were down-regulated, particularly for CXCL12 (SDF-1). NGFR was significantly down-regulated, consistent with flow cytometry data, as well as some, but not all, osteogenesis-related molecules. Interestingly, all 3 immature transcripts were down-regulated potentially suggesting some loss of “immaturity” following cultivation.

Transcriptome of in vitro MSCs

The transcriptomes of osteogenically-differentiated MSCs (oMSCs) and in vitro MSCs cultured using standardised PurStem procedures were compared with those of skin fibroblasts (control mesenchymal lineage cells that lack robust tri-potentiality and are poorly osteogenic) and endothelial cells (non-mesenchymal lineage cells). The results established close lineage relationships between in vitro MSCs, oMSCs and fibroblasts and provided proof-of-principle that oMSCs generated using standardised PurStem procedures had a characteristic transcriptional profile. These experiments also confirmed that the chosen gene microarray methodology was powerful enough to detect differences between primary cells of disparate cell types.

In addition, the transcriptomes of oMSCs and cultured MSCs were compared with in vivo MSCs. As expected, all adipose-lineage specific transcripts (PPARG, LPL and FABP4) were expressed in in vivo MSCs at high levels, confirming their potential for adipogenesis. Their expression in cultured MSCs was lower, potentially reflecting culture-adaptation phenomena, and FABP4 expression in particular was further down-regulated in oMSCs. These data confirmed that FABP4 was the best adipogenic marker for MSCs.

Major osteogenic markers (RUNX2, SPP1, SPARC) were also highly-expressed in in vivo MSCs. Of the osteogenic markers tested, only SPP1 was down-regulated in cultured MSCs, in contrast to adipogenic markers, which were all down-regulated. SPP1 was additionally up-regulated in oMSCs, indicating that it is the best mature osteogenic marker from the markers tested. Importantly, the levels of all osteogenic markers in in vivo MSCs were as high as those in oMSCs when compared to cultured MSCs, potentially reflecting their derivation from the native bone rich environment and also indicating their high-level of osteogenic commitment. Overall, these data indicated that oMSCs generated using standardised PurStem procedures more closely resemble in vivo MSCs than cultured MSCs. On the other hand, MSCs cultured under standardised PurStem conditions were clearly capable of differentiating to oMSCs, given the necessary stimuli, thus confirming the potential of cultured MSCs to acquire the phenotypic characteristics of in vivo MSCs.

The chondrogenesis-related transcription factor SOX9 was expressed in in vivo MSCs but mature markers COL2A1 and EPYC were absent (data not shown), suggesting that in vivo MSCs were “ready for”, but not actively engaged in, chondrogenic differentiation. The lack of mature chondrogenic protein marker expression in in vivo MSCs is consistent with the fact that, in adults, chondrogenesis occurs only following trauma.

Validation of novel receptor expression by flow cytometry

Flow cytometry was used to validate the expression of selected transcriptome and receptome (WP3) receptors in cultured and in vivo MSCs. In accordance with the transcriptional profile, protein expression of both TGFbR2 and TGFbR3 was significantly down-regulated in cultured MSCs compared to in vivo MSCs. TGFbR3 expression was almost completely absent at Passages 1 and 3. TGFbR2 expression was also down-regulated, although the decline was more gradual, decreasing with subsequent culture. As observed with the transcriptional data, FGFR1 expression was maintained at similar levels in cultured and in vivo cells regardless of the passage number. However, the protein expression level of both FGFR2 and FGFR3 was reduced in cultured MSCs compared to in vivo MSCs.

Despite observing reduced FZD1 and FZD4 transcription in cultured MSCs compared to in vivo MSCs, minimal protein expression was detected for either receptor. Frizzled 9 (CD349) expression was detected on in vivo MSCs and its expression was found to increase during early culture. Whilst this does not entirely correlate with changes in transcription, this may be an effect of post-transcriptional regulation. Since cultured MSCs lose their osteogenic potential with time, Frizzled 9 may be a useful marker to predict osteogenicity (WP1).

Validation of receptor expression on osteogenic cells
Cultured MSCs at passage 2 were seeded and cultured for 2 weeks using either standardised PurStem or osteogenic conditions. Surface markers were then determined using flow cytometry.

Of the markers tested, lineage specific markers, such as CD14, CD19, CD31, CD34, CD45 and HLA-DR, were negative in both the osteogenically differentiated cells and undifferentiated control cells. Most of the novel receptors tested, such as TGFbR2, FGFR1 and Frizzled 4, showed little difference in expression levels between the two cell type. Of the ‘positive’ markers tested, no differences were observed in the expression levels of CD73, CD90 or CD105 for either cell type. These results demonstrate the limitations of using these ‘positive’ MSC markers as they may robustly identify MSCs they do not reflect the functional or differentiation state of the cells.

Frizzled 9 (CD349) showed the largest difference in expression levels between undifferentiated and osteogenically differentiated cells. Although there was a clear pattern of reduced Frizzled 9 expression on osteogenically differentiated cells, the difference was not statistically significant. This is likely to be a reflection of the low numbers of sample tested in this assay. The fact that CD349 appears to be down-regulated in ‘aged’ MSCs and in osteogenically differentiated cells is interesting. Further research to determine whether this marker is also associated with adipogenic differentiation could provide further evidence that CD349 could serve as a potential marker of true immature and multipotential MSCs.

1.3.3 MSC Receptome (WP3)

Why characterise the MSC Receptome?

Characterisation of the MSC Receptome is important to enable reliable identification of MSCs. To date, no single marker or unique set of markers has been identified that distinguishes a pure, native MSC population from other cell types with intrinsic MSC qualities, such as stem cell precursors, endothelial cells or epithelial cells. In 2006, the International Society for Cellular Therapy (ISCT) published a position paper proposing a minimal set of standards to define human MSCs. In addition to plastic adhesion properties and multipotent differentiation potential, the ISCT proposed a specific surface antigen expression profile that should include, at least: ?95% positive for CD105 (SH2), CD73 (SH3/4) and CD90 (Thy-1) and ? 2% expression of CD45, CD34, CD14/CD11b, CD79?/CD19 and HLA-DR. However, this list is by no means exhaustive. Additional markers that have been associated with MSC identification include CD44, CD133, CD49a, LNGFR, CD10, CD13, BMPRIA and STRO-1 antigen molecule, in addition to adhesion molecules VCAM-1, ALCAM, ICAM-1 and CD29.

PurStem researchers sought to establish a novel database of receptors that would provide:
• promising leads for growth factor combinations or coating procedures that would control MSC adhesion, proliferation and differentiation in serum-free media
• potential new immuno-reagents that could be used to strengthen the clinical release criteria for MSCs
• potential new methods for direct isolation of MSCs from tissue based on specific immunoselection rather than adherence.

PurStem Results

High-throughput fluorescence-activated cell sorting (FACS) analysis

Bone marrow from three donors was collected and processed at NUIG according to the standardised PurStem procedures for isolation, sub-culture and cryopreservation (PS cells). Cells were also cultured under standardised PurStem conditions, but at a lower plating density (LD cells) reflective of an optimal GMP plating density. Additionally, MSCs were cultured under reduced serum conditions developed by PurStem partner CUNI in WP4 (CUNI cells). Samples from the three conditions were cryopreserved and shipped to Becton Dickinson (BD) for high-throughput FACS analysis.

Expansion of the PS, LD and CUNI cells showed significant donor-to-donor variability in the number of colonies formed, but each donor maintained similar population doublings within a culture condition. Cells from all donors cultured in PS and LD conditions had comparable population doublings and cell yields indicating comparability between donors. CUNI cells demonstrated more donor-to-donor variability. The high-throughput FACS analysis by BD was consistent with these observations. The LD and PS heat maps closely resembled each other, while the CUNI expression of antigens varied slightly.

24 markers that were positively expressed on MSCs in all culture conditions were identified and included the positive ISCT markers CD73, CD90 and CD105 among others. The ISCT negative markers CD45, CD34, CD14/CD11b and CD79?/CD19 were not detected, confirming the validity of the BD results. A further 33 markers that are highly expressed in PS and LD cultured cells were identified. The markers identified, including the ISCT positive markers, are not specific only to MSCs. Therefore, the PurStem team carried out a target validation to identify markers characteristic for MSCs that complement the ISCT markers, as outlined below.

Target validation - in vivo MSCs from bone marrow

Surface marker analysis was performed on in vivo MSCs extracted enzymatically from marrow filling trabecular bone cavities. Trabecular bone is a fairly abundant MSC source that can be obtained easily during fracture surgery or from discarded femoral heads removed during hip replacement surgery for osteoarthritis (OA). In vivo MSCs from hip OA femoral heads were compared with early-passage cultured MSCs, grown under standardised PurStem conditions from matched donors. The surface marker results were similar to those observed for MSCs from non-osteoarthritic donors. Amongst the markers studied, CD146 was the most strongly expressed on in vivo OA MSCs, followed by TGFBR2. After the first passage, a general trend for a decline in marker expression was observed for FGFR2, TGFBR2 and TGFBR3 but not for CD146, CD106 or FGFR1, which was consistent with normal donor data. FGFR1 expression was stable during cultivation, whereas TGFBR3 was down-regulated at both the RNA and protein level.

Both in vivo and cultured OA MSCs had lower expression levels for selected receptome markers and hence the decrease could not be attributed to the enzymatic digestion procedure that was performed to extract in vivo MSCs. These results provided the first indication that the expression of surface markers in OA MSCs is different than normal MSCs, which may have some functional relevance to the disease.

ECM-Adhesion Molecules and Growth Factors for Serum-Free Culture

In order to support the development of serum-free culture conditions (WP4), highly expressed surface markers that are also extracellular matric (ECM)-adhesion molecules and growth factor receptors were identified. Nine ECM binding proteins were highly expressed on MSCs cultured under CUNI, LD or PS conditions. Primarily, these cell surface markers were integrin isoforms that bind fibronectin, although CD61 and CD44, which bind vitronectin and hyaluronic acid, were also highly expressed. These results indicated that coating culture flasks with fibronectin when isolating MSCs in serum-free conditions should enhance cellular adhesion. Although there were several growth factor receptors expressed, the majority of the expressed markers were associated with inflammation (IL receptors, TNF receptors) or cellular differentiation (TGF-beta). However, those receptors associated with pro-proliferative signalling (EGF, PDGF receptors, chemokines) were identified as potentially useful for serum-free culture expansion.

Effect of culture conditions on MSC receptome

Further analysis was conducted to characterise the effect of culture conditions on protein expression. MSCs were cultured under four culture conditions (PS, LD and CUNI cells as above, plus REMEDI cells cultured under standardised PurStem conditions without FGF supplementation). Unique proteins were expressed in each culture condition. CUNI cells expressed fewer proteins in general while the PS, REMEDI and LD protein expression levels were comparable as expected due to their similar culture conditions.

The expression of several ECM binding proteins on the MSC surface was confirmed, thereby validating the BD FACS data. Tissue culture plastic was coated with the complementing ECM molecules (fibronectin, vitronectin or hyaluronic acid) and demonstrated equivalent or enhanced MSC binding as compared to uncoated plastic, validating the expression of these ECM binding proteins on the cell. Internal signalling through these receptors was confirmed through western blotting.

Combinatorial exposure of MSCs in serum free conditions to fibronectin and the growth factor EGF resulted in cell attachment that was enhanced as compared to serum-containing conditions, suggesting a combination of these factors for further exploration in WP4. The presence of several growth factor receptors identified by FACS analysis on MSCs was also confirmed. MSCs cultured in the presence of these factors in the absence of serum enabled cellular attachment and proliferation at rates comparable to serum containing medium. Western blotting was used to similarly confirm the presence of several growth factor receptors in MSCs as suggested by BD FACS analysis.

Effect of long term culturing on MSC receptome

MSC cultures from 3 donors were cultured long-term under standardised PurStem conditions and marker expression was tested at passages 1, 3, 5 and 7. The panel of markers analysed was slightly modified from that used above to exclude negative or very low expressing markers (CD220, CD221, Pref1 and CDCP1) and to include functionally-relevant markers, such as CD106 and CD146. CD146 was expressed on the overwhelming majority of cells in all three MSC cultures (>98% cells positive) and no obvious decline was observed with passaging.

CD146 has been recently described as a marker of self-renewing osteoprogenitors in the marrow. UL historic data on MSCs cultured in NH media (standard MSC expansion media from Miltenyi Biotec) indicated the loss of CD146 during long-term cultivation. The fact that CD146 was stably expressed during long-term cultivation under standardised PurStem conditions indicates better preservation of osteoprogenitor activity under PurStem conditions. However as reported in WP1, CD146 expression was highly variable between partners. CD106 is a vascular cell adhesion molecule (VCAM) whose expression has also been shown to decline significantly following cultivation in NH media. Under standardised PurStem conditions, CD106 expression remained stable, although higher donor-to-donor variation was observed for CD106 than CD146. These results indicate the potential superiority of standardised Purstem media over NH media.

1.3.4 Serum-Free MSC Culture (WP4)

Why is serum-free culture needed?

Clinical application of MSCs will require expansion and differentiation of primary MSCs in order to obtain sufficient therapeutic cell numbers as it is not possible to get enough MSCs for clinical use during a single expansion. Usually, the cells have to be passaged 1-2 times in medium that contains animal-derived reagents, which increases the risk of contamination of the cell culture and also increases the time from the bone marrow harvest to the final production of the cellular treatment to 4-6 weeks.

Traditionally, MSC culture has been performed using bovine serum containing medium due to its high levels of growth-stimulatory factors and low levels of growth-inhibitory factors. However, there are a number of risks associated with using animal derived products for the culture of therapeutic cells, including prion, viral or zoonose contamination and patient immune reaction. In addition, there is a need for serum-free culture conditions for quality control of experiments between laboratories, as the variability of bovine serum can affect the reproducibility of results. For these reasons, one of the goals of PurStem was to develop serum-free media conditions to support the manufacture of safe, GMP-compliant stem cell therapies.

Commercially available serum-free media

Currently, there are commercially available serum-free media. Mesencult-ACF from Stem Cell Technologies and Thera Peak from Lonza were tested during the PurStem project. Unfortunately, the use of current serum-free conditions in culturing human MSCs selects for a subpopulation of cells that can survive serum deprivation and continue proliferating. A new complement of medium, growth factors, cytokines, etc. must therefore be developed specifically ensure consistent MSC cell surface receptor expression yielding uniform expansion of the MSC population in culture.

PurStem Results

Comparison of PurStem medium and commercial serum-free media

Colony forming assays and expansion cultures were set up in standardised PurStem (PS) medium, which contains serum, low serum media (2% fetal calf serum (FCS)) and two commercially available serum-free media (Mesencult-ACF and Thera Peak). The cellular morphologies of the cultures were compared after 14 and 30 days.

MSC colonies were successfully cultured in commercially available serum-free media (Mesencult-ACF, Thera Peak) over the short term (14 days). After 14 days, the colonies observed in the standardised PurStem medium were large, with densely packed cells, while those from the Mesencult-ACF medium were small, containing fewer and more loosely packed cells. After 30 days, the cells in serum free media failed to proliferate and morphological changes were observed while the cells in PurStem medium were confluent. In Therapeak medium, spheroids were visible following the first passage. Despite the same initial seed concentration, the Mesencult-ACF cultures contained only 25% of the number of cells of the PurStem medium cultures. Although our results demonstrated that it is possible to culture MSCs in commercially available serum-free media, these media were not suitable for sustaining MSCs in long term culture. Therefore, improved serum-free media conditions are required to culture MSCs for use in clinical therapy.

Effect of Growth Factor Combinations and ECM Molecules on MSC Viability and Proliferation

Using data from WP3 and WP6, PurStem researchers compiled a list of relevant factors with the potential enhance MSC cell attachment and proliferation, including mitogenic factors and ECM molecules, for screening. A serum-free medium from the literature developed for human articular chondrocytes, containing a cocktail of recombinant growth factors to drive cells towards osteogenic or chondrogenic differentiation was also tested. MSC cells cultured under these serum-free conditions were elongated and fibroblast-like, similar to those cultured in PS media, and retained the ability to undergo osteogenic differentiation.

MSC attachment after 24 h was examined using a range of ECM molecules, including vitronectin, fibronectin, laminin and hyluronic acid. For MSC isolated in the presence of serum, vitronectin seemed to result in the best initial cell attachment without altering cell morphology. However, after 6 days, MSC proliferation was greatest on fibronectin coated plates. Further study showed that fibronectin promotes MSC attachment by phosphorylation of protein FAK after 90 minutes.

MSC proliferation in the presence of growth factors on fibronectin coated plates in serum free media was examined. The growth factors examined included PDGF-aa, bFGF, FGF4, IGF2, EGF and GM-CSF. Initial results suggested that bFGF and FGF4 play a role in proliferation. Further optimisation of the serum-free media formulation was conducted, and western blot analysis revealed that the presence of EGF in serum-free media promoted MSC attachment, but did not promote cell proliferation. There was a statistically significant increase in cell proliferation for media with PDGFaa/FGF2 or FGF4/FGF2 or FGF4 compared to serum free media. There was no statistically significant difference between MSC cultured in FGF2/FGF4 and FGF4 alone.

Using colony forming assays and xCELLigence technology which allows monitoring of cellular events in real time, the team at CUNI tested several combinations of growth factors/supplements and performed high throughput analysis of proliferation and viability. Basic combinations of growth factors/supplements with serum-free media were tested using high throughput analysis of proliferation, viability assessment, phenotyping and pulsed electric field (PEf) measurements. Cells cultivated in high concentrations of PDGF-BB were more numerous but less stable during long term cultivation. Moreover CD146 expression decreased as the amount of PDGF-BB used was increased. A combination of EGF (10 ng/ml) and ITS (10 ul/ml) with the addition of HAS (0.125%) yielded an optimum basal serum-free media composition.

Surface markers of MSCs cultured in serum-free conditions

Based on the results of from WP6, additional supplements were tested by direct plating in PS media, low serum (2% or 1% serum) and serum-free media. To assess the effect of the supplements, CD146 and CD271 expression levels on the cells were examined at frequent time points.

Cells grown under standardised PurStem conditions (10% serum) had a standard MSC phenotype. Cells grown in other media conditions had an MSC phenotype (expressing CD73, CD90, etc. and lacking expression of lineage markers) but did differ in the expression levels of some markers. The most prominent difference was a decrease in expression of CD105 and CD146 for the three media tested and an increased expression of CD271 for Medium A when compared to cells grown in standardised Purstem media.
MSCs were cultivated in special culture dishes (PureCoat Amine and Carboxyl, BD) to determine if cells could be cultured in serum-free media, supplemented with EGF, without plate coating to meet GMP requirements. PureCoat Amine dishes were suitable for serum-free culture, while PureCoat Carboxyl was unsuitable.

Differentiation potential of MSCs cultured in serum-free conditions
We examined whether the MSCs grown under the optimised PurStem serum-free conditions differentiated towards mesenchymal lineages. The serum-free medium contains a cocktail of recombinant growth factors specifically used to drive cells towards osteogenic or chondrogenic differentiation, in association with other components involved in normal cell homeostasis and metabolism. We compared the osteo- and chondrogenic potential of serum free expanded cells with the potential of cells grown in serum containing media.
Results indicated that MSCs cultured under the PurStem serum-free conditions show an elongated and fibroblast-like phenotype. Nonetheless, they maintain the ability to undergo osteogenic differentiation.

Serum-free cryopreservation

In classic serum-containing freezing media, serum exerts a protective function for the cells. BIOFREEZE (Biochrom AG, Germany) is a commercially available medium free of animal-derived substances that is also free of genetically modified organisms. MSC (passage 1) cells cultivated in serum-free medium supplemented with EGF were cryopreserved using the standardised PurStem cryopreservation procedure (WP1) and BIOFREEZE for comparison. There was no significant difference between the number of cells frozen and number of thawed viable cells. The number of recovered cells at passage 2 and viability of the samples immediately after thawing was significantly better for the standardised PurStem procedure than the BIOFREEZE procedure. However, after one consecutive passage (Passage 2), the viability of both protocols was comparable.

PurStem serum-free media were also developed at NUIG. These second generation media were also capable of promoting isolation and expansion of tri-potential MSCs to passage 7 with characteristic surface marker expression. Growth to passage 9 was significantly improved over culture in serum-containing medium and serum-free MSCs had significantly higher surface expression of CD271 to passage 6 and CD146 to passage 9.

1.3.5 In Vivo Potency (WP5)

Why measure MSC osteogenesis?

MSCs were first identified in marrow on the basis of their osteogenic activity and are tested for their ability to deposit a mineral-rich matrix. More importantly, bone tissue engineering is one of the attractive clinical applications of MSCs in humans. Although in vitro differentiation assays are important in terms of release criteria it is necessary to establish the potential of MSCs to contribute to tissue formation in vivo. For these reasons, one of the goals of PurStem was to assess the in vivo osteogenic potency of in vitro MSCs derived using the PurStem standardised methods.

How is in vivo osteogenesis measured?

The ceramic cube assay is a semi-quantitative test of the ability of culture-expanded cells to produce bone or cartilage in an osteo-conductive environment in vivo. In particular, when combined with mineral containing tri-dimensional scaffolds, MSCs form a primary bone tissue, which is highly vascularized and colonized by host hematopoetic marrow. In this system, the ceramic provides a tri-dimensional structure, a cell adhesion site and may act as a primer for the formation of new bone matrix. In addition, the MSCs differentiate into osteoblasts and deposit extracellular matrix on the ceramic surface. The ectopically formed tissue may be quantified by histological techniques.
Skelite™ is a resorbable bioceramic based on silicon stabilized tricalcium phosphate (Si-TCP). In addition to stimulating the bone healing process, Skelite™ is gradually removed by bone-resorbing cells. This permits natural remodelling and replacement with the patient's own bone. Skelite™ could be used in a wide spectrum of bone graft applications, including repair of bone defects in the spine and pelvis.

PurStem Results

Skelite ectopic cube assay

MSCs were cultured by all four partners using standardised PurStem procedures (WP1). Cryopreserved MSCs were shipped to UNIGE for osteogenic evaluation using an ectopic cube assay with a ceramic cube system (Skelite™). PurStem cells at passage 2 from each of the partner laboratories were implanted in immuno-compromised nude mice using Skelite™ cubes. The scaffolds were recovered after 4 and 8 weeks for histological analysis. Sheep MSCs implanted with Skelite™ cubes were used as a control.

After 8 weeks, we observed a small amount of bone formation in scaffolds implanted with cells obtained from all four partners. However, the bone formation from cultured MSCs was poor compared to the positive control (sheep MSCs). The osteogenic potency of in vivo MSCs isolated by UL in WP2 was also determined. The number of cells obtained using this approach is very low compared to the number of cultured MSCs obtained using standardised PurStem expansion procedures (WP1). Therefore, the in vivo MSCs were implanted with mytomicin C treated fibroblasts (1:500 ratio of MSCs to fibroblasts). Cartilage and bone formation were not observed after 8 weeks but fibrous tissue, adipose tissue and vessel formation were observed.

The osteogenic capacity of MSCs cultured under serum-free conditions was determined using cells prepared by CUNI (WP4). For comparison, MSCs cultured from the same donor using standardised PurStem conditions (WP1) were also implanted. No bone formation was observed for MSCs cultured under these serum-free conditions.

MSC isolated, expanded and cultured in 2nd Generation PurStem serum-free media, developed at NUIG, showed a good bone deposition in vivo compared with the cells cultured in serum containing medium.

Third generation osteoinductive scaffold

After initiation of the PurStem project, a new 3rd generation osteoinductive scaffold became available. The new scaffold consists of a ceramic of biphasic calcium phosphate with a different percentage of hydroxypatite and tricalcium phosphate beta than Skelite™. This new synthetic biomaterial was loaded with MSCs previously tested on the Skelite™ cube and the same osteogenic protocol was followed. Unlike the results with Skelite™, the amount and quality of bone formed with implantation of MSCs on the new biomaterial was comparable to the positive control (sheep MSCs).

1.3.6 Novel Antibodies for MSC Surface Recognition (WP6)

Why do we need antibodies for MSC surface recognition?

MSCs represent a tiny percentage of human adult bone marrow (estimated at 1 in 100,000 mononuclear cells). This rarity combined with the need for significant numbers of MSCs for clinical treatments and the heterogeneity of MSCs creates a need for methods using standardised isolation reagents to purify and characterise MSCs. PurStem researchers identified novel antibodies as a way to enhance the current state of the art for isolation and characterisation of MSCs.

How do we currently distinguish MSCs from other cell types?

There is little consensus among scientists regarding a method for the isolation of MSCs from human tissue. There is some consensus on the defining characteristics of MSCs, based on a recent position paper from the International Society for Cellular Therapy (ISCT). However, this definition is based on a somewhat slender list of biological attributes and, in the words of the authors, represents research criteria that “should not be confused with release specifications for clinical studies”. The authors further indicate that the criteria are based on “the best currently available data” and “future research will probably mandate a revision”. The lack of agreed clinical release specifications is a serious impediment to progress in assessing the therapeutic potential of MSCs in humans. There is a risk that current clinical studies will be rendered useless by the lack of consensus on the definition of MSCs and the lack of standardised isolation protocols.

The ISCT position paper gives adherence to plastic, specific surface marker expression and multipotent differentiation potential as the criteria for defining MSCs. No single criterion in this list is specific to MSCs alone. Indeed, it is probable that other cells types share ALL of the MSC-selected criteria outlined in this position paper. Certainly this merely reflects the inadequacy of our understanding of the phenotype, genotype and plasticity of MSCs. There is a conundrum in stem cell technology that is highlighted in the application of MSCs more than any other stem cell type:

• the absence of a rigorous and foolproof set of criteria for a standard definition of an MSC means that universal standards of preparation are essential
• the absence of universal standards of preparation means that release criteria must be solid and cell-specific.

Stem cell researchers have not yet to reach either standard. For these reasons, PurStem sought to identify novel surface antibodies that could be used in GMP-grade environments for the clinical isolation of MSC from human tissue for therapeutic applications and generate new standards for isolating and defining MSCs that would support industrialisation of stem cell technology.

PurStem Results

Microarray profiles and bioinformatics analysis

MSCs from human bone marrow were isolated and expanded using standardised PurStem procedures (WP1). For analysis of particular genes using Affymetrix® microarrays, it was necessary to use alternate protocols - bone marrow aspirate was Ficoll purified and FGF2 was not utilised in MSC culture. Flow cytometry was used to confirm expression of surface markers.

Affymetrix® data was generated from cultured human MSCs (Passage 2), five fibroblast cell lines, CD34-positive cells, bone marrow mononuclear cells (BMMNCs), endothelial cells, and smooth muscle cells. Bioinformatic comparative cluster analysis of the data identified at least 1008 protein coding RNAs that discriminated MSCs from fibroblasts.

Based on the targets identified by the Affymetrix® experiments, initial flow cytometry experiments were carried out with the first antibodies available that reacted positively with human MSCs. These results reduced our list of potential antibodies from 12 to 5. Note that some of the proteins identified, such as PDGFRa, ITGB5, and ITGA10, have been previously reported as MSC markers, thus validating our approach. However, the focus was to identify novel antibodies that bind proteins on the surface of human MSC. One of the potential surface markers had not been reported in the literature or by PurStem partners. A commercially available antibody that binds to this marker was identified and has been used in combination with the UL approach for isolating in vivo MSCs to isolate a sub-population of MSCs at high purity. Based on PurStem results, a United Kingdom patent application has been filed by Orbsen Therapeutics Limited.

1.3.7 Production of “germ-free” MSC surface recognition antibodies (WP7)

Why are “germ-free” MSC surface recognition antibodies needed?

Given the low percentage of MSCs present in adult bone marrow (approximately 1 in 100,000) compounded by the need for significant numbers of MSCs for clinical treatments, in vitro culturing of MSCs is necessary to obtain sufficient numbers of cells for clinical applications. The success of culturing a significant population of MSCs can be enhanced by either initial purification of MSCs from bone marrow samples or by purification of the cultured MSCs. Both of these approaches can be achieved using antibody-based affinity purification of the MSCs. In this context, large-scale production of “germ-free” antibodies is necessary for commercial scale, GMP-compliant purification of MSCs for clinical applications. For these reasons, one of the goals of PurStem was to develop GMP-compliant transgenic chicken lines, capable of expressing MSC-specific antibody fragments in hen egg white.

How are antibody fragments generated?

Antibody fragments were generated by immunising chickens with human BM derived MSCs. Bio-panning experiments were carried out to identify antibody fragment clones, which were highly specific for MSCs. One of these clones, TMSC3, was selective for MSCs over other bone marrow derived cells, making it an ideal candidate for an MSC affinity purification reagent. As TMSC3 was derived from a chicken antibody, it was decided that expression of this antibody on a large enough scale to yield sufficient protein for use in a GMP-compliant purification procedure would be best carried out in a chicken expression model to maintain essential binding characteristics as possible.

How are chickens used to express transgenic proteins?

Since the advent of the first successful cloning by nuclear transfer at the Roslin Institute in 1986 the prospects of generating animals to express foreign proteins for commercial scale production have been evident. The Roslin Institute has established various methods, including nuclear transfer, for the generation of transgenic chickens expressing foreign proteins after insertion of foreign DNA into the animals. This technology has been refined and is used to express bioactive proteins at high yields.

PurStem Results

Generation of lentiviral construct and packaging of retroviral construct

Previous DNA constructs used for the generation of transgenic chickens required multiple cloning steps. The greater the number of cloning steps the greater the risk of errors and failure. Therefore, PurStem partner Ovagen adopted an approach using novel DNA synthesis technology to construct the required DNA insert de novo from sequence data, thereby eliminating the need for extensive cloning. As it was not possible to include the ovalbumin promoter as part of this de novo construct, the final construct was cloned in two steps. The final construct was amplified in bacteria and purified prior to assembly into a full retroviral package for chicken embryo infection.

Generation of transgenic chickens

Embryos were injected with the TMSC3 construct and transgenic chickens were identified after hatching (G0). Male transgenic cockerels were grown to sexual maturity and semen was screened for the transgene. Three G0 male chickens were used to generate a G1 population. A G1 male was selected to inseminate stock female chickens to generate the G2 generation, which will form the main commercial TMSC3-producing flock. Transgenic G1 hens were selected to produce pilot batches of TMSC3.

Isolation, purification and characterisation of TMSC3

TMSC3 was isolated from the egg whites for characterisation. Chicken egg whites are highly viscous due to the presence of a protein called ovomucin. The viscosity was reduced using a specialised precipitation procedure optimised by PurStem partner Ovagen. Following viscosity reduction, TMSC3 was purified from the egg whites using Immobilised Metal Affinity Chromatography (IMAC). After purification, the protein was 90% pure and the yield was approximately 2.3 mg/ml, ~50mg total TMSC3).

The above procedures were carried out on the TMSC3 that was dissolved within the egg white. However, inspection of freshly laid and stored eggs revealed a coarse white precipitate in the egg white, which is not present in the egg white of age matched control chickens. This may indicate a larger pool of TMSC3, which could potentially be separated from the egg white by centrifugation.

The protein precipitate was isolated from eggs that had been stored for 10 days. Analysis showed the presence of a protein the size of TMSC3 that represented 60% of the total protein (approximately 10 mg/ml or 350 mg/egg). These results suggest that it may be possible to harvest very pure material from egg whites without the need for time consuming IMAC purification, which in turn will have a beneficial effect on commercial scale production. This level of protein production is comparable with state of the art in vitro systems currently used for the production of recombinant proteins.

1.3.8 GMP Compliance (WP8)

Why is GMP compliance needed for MSC isolation and expansion?

There has been relatively little progress in the development of new culture technologies for the large-scale manufacture of mesenchymal stem cells (MSCs). There is a strong possibility that this limited ability to produce stem cells will result in delays to the translation of new therapies to the clinic. The current state of the art has several weaknesses, in that, there are no standards for characterisation, isolation or identification of MSCs from any tissue, nor are there standard protocols for differentiation of MSCs to various lineages. Additionally, surface markers used for MSC characterization lack specificity and cryopreservation protocols are not standardised. Critically, current production methods for MSC require the use of animal products with major contaminant implications. One of the goals of PurStem was to identify GMP-compliant raw materials, develop an SOP for small scale serum-free media production, and validate analytical assays for use as Quality Control release assays for the production of GMP-compliant cells for clinical applications.

What are the requirements for GMP production of therapeutic MSCs?

PurStem aimed to prepare for the advanced future manufacturing demand for MSCs in support of the industrialisation of stem cell technology. For clinical safety, all cell based medicinal products must comply with EU safety requirements. GMP requires that raw materials and other consumables used in the production and quality control analysis of products are of GMP grade to ensure the consistency of release results and patient safety. Standard operating procedures (SOPs) and validation of quality control release assays are also essential for GMP compliance with EU regulations.

PurStem Results

GMP-compliant raw materials and serum free media production

PurStem partner NUIG generated a list of raw materials that comply with GMP requirements, prepared a report outlining regulatory requirements and generated an SOP for the small-scale production of serum-free media as reference documents.

Validation of quality control assays for MSCs

Current legislation governing the manufacture of human medicinal products, in particular advanced therapy medicinal products, within the European Union dictates a number of specific quality criteria for the product outputs of a manufacturing process. In addition, certain other quality criteria as regards product identity or process-related impurities are defined from the particulars of the manufactured products specification details. In this context, Quality Control analytical tests for in vitro culture-expanded MSCs should assess the following properties: total cell density, morphology and viability assessment, MSC cell surface biomarker profiling, differentiation capability (osteogenesis, chondrogenesis, adipogenesis), karyotype analysis, sterility assessment, bacterial endotoxin assessment, mycoplasma assessment and removal or determination of residual amounts of process-related impurities.

Using published guidances, a cell surface marker profile characteristic with MSCs will be used as the basis of an analytical assay to determine the degree of purity (MSC population) of the final manufactured product. The panel of antibodies to profile cell surfaces for biomarkers characteristic of MSCs (MSC Receptome) as presented below, was developed based on existing markers and novel markers identified by Purstem WPs 3, 6 and 7. Before finalization of the panel, an interim quality control analytical assay using existing cell surface markers was validated as per GMP practice, using MSCs produced under the standardised PurStem protocol reported in WP1.

Potential Impact:
Research and translational impact

Pre-clinical and clinical trials have demonstrated the enormous potential of stem cell therapies for the treatment of a wide range of diseases, including myocardial infarction, osteoarthritis and inflammatory diseases. In order to fully exploit the potential of stem cell therapies and meet current regulatory requirements, standardised methods for isolating, characterising and manufacturing stem cells beyond the current state of the art are required. PurStem has advanced the state of the art in adult human MSC isolation, characterisation, expansion and biology by:
• Developing and validating new standardised GMP-compliant procedures, reagents and media formulations for stem cell isolation, expansion, differentiation and banking
• Characterising the receptome of both in vivo and cultured MSCs
• Developing novel antibodies against MSC surface markers to specifically identify and ultimately isolate MSCs
• Developing “rational” animal product-free / serum-free media
• Establishing novel technology for “germ-free” expression of novel antibodies.

Standardised PurStem procedures for the isolation, expansion, differentiation and cryopreservation (banking) of MSCs will contribute to the optimisation of GMP manufacturing and banking of MSCs for use in clinical trials initially and ultimately as a commercial product. As regulatory authorities approve more stem cell therapies for use in clinical medicine, it is anticipated that there will be a gap between supply and demand. The results of PurStem will facilitate large-scale commercial manufacturing of GMP-grade MSCs to fill this anticipated shortfall.

PurStem has also advanced our basic understanding of MSC biology by defining the surface receptome. PurStem researchers have shown that the in vivo MSC receptome is highly-tuned to respond to stimulation by osteogenic ligands. Expression of many, but not all, osteogenic receptors was found to be down-regulated in MSCs cultured under standardised PurStem conditions. Down-regulation of osteogenic receptors in culture tended to preserve MSC osteogenesis better than stromal supportive function or adipogenesis at the transcript level. Furthermore, analysis of the receptome indicated that Frizzled 9 may serve as the best potency and culture quality indicator for MSCs cultured using standardised PurStem conditions. These results have important implications for potential uses of MSCs in clinical therapies as they provide researchers with the means to ensure the production of potent, high quality MSCs capable of osteogenic differentiation for therapeutic applications.

The work of PurStem will contribute to the on-going efforts of the International Society for Cellular Therapy to establish internationally-recognised standard operating procedures for stem cell research. For example, the ISCT position statement8 indicates that MSCs must be positive for the expression of CD73, CD90 and CD105. Analysis of the MSC receptome by PurStem researchers showed that CD105 expression for cells cultured under standardised PurStem conditions was variable between laboratories, indicating that this is not the most suitable marker for defining the MSC phenotype and may in part contribute to the disparity of MSC phenotypes reported in the literature. Furthermore, CD73 and CD105 do not indicate proliferative or osteogenic capacity for cultured MSCs.

Although clinical trials were not carried out as part of PurStem, the project included a strong translational focus:
• Pre-clinical model of bone formation: The capacity of in vivo and cultured MSCs obtained using the standardised PurStem procedures to form bone and cartilage in vivo when implanted with ceramic scaffolds was investigated using a pre-clinical mouse model. Implantation of MSCs with a novel 3rd generation scaffold yielded promising bone formation results.
• GMP-compliant procedures: Patient safety and GMP compliance were accounted for when developing new reagents and standard operating procedures so that MSCs manufacturing processes based on the PurStem approach will be suitable for both clinical and commercial use. GMP-compliant raw materials were also sourced for progressing the state-of-the-art in the production of MSCs in large quantities. An SOP was developed to standardise the small-scale production of serum free media for use in the production of GMP-compliant MSCs and a GMP-compliant, analytical quality control release assay for MSCs was validated.
• Germ-free, GMP-compliant protein manufacturing: A significant amount of effort was dedicated to developing technology for manufacturing germ-free, GMP-compliant proteins for the formulation of growth-factor supplemented serum-free media and MSC affinity purification reagents. Proof-of-concept experiments using the affinity purification reagent TMSC3 showed that transgenic chickens could be used to manufacture GMP-compliant proteins at high yields.

PurStem dissemination efforts, discussed in detail below, targeted other stem cell researchers in the partner countries (Ireland, the UK, Italy and the Czech Republic) and globally in order to achieve maximum research and translational impact. Additionally, GMP stem cell manufacturers, clinicians involved in clinical trial design and regulatory authorities were key dissemination audiences.

Socio-economic impact

The area of stem cell therapeutics is expected to lead to treatment of diseases for which there are currently no effective treatment options, as well as contribute to tissue-engineering of new tissues or organs for replacement purposes. Promising disease areas for MSC applications currently include:
• cardiovascular applications, e.g, repair after myocardial infarction
• inflammatory conditions such as Graft vs Host and Crohn’s disease and
• osteoarthritis and other orthopaedic problems.

A key issue for the application of MSCs to these and other conditions is the lack of standardised procedures for the generation of well-characterised, GMP-compliant MSCs in industrial quantities. PurStem has advanced the state of the art in this area and in this way will facilitate the development and clinical application of stem cell therapies for these diseases.

Members of the PurStem consortium are experts in the area of arthritis and bone repair. Hence, while many of the PurStem results can be generally applied to the manufacture of stem cells for a range of diseases, there was also a particular focus on the ability of MSCs to generate bone, cartilage and adipose tissue due to the therapeutic potential of these regenerative abilities in arthritis and bone repair. Osteoarthritis is a disease associated with ageing and is the most common of the rheumatic diseases. In 2005, the European League against Rheumatism (EULAR) estimated that approx. 100 million people in Europe suffer from rheumatic diseases, accounting for half of all chronic conditions in persons aged over 65 (www.boneandjointdecade.org) with the quality of life of about 7.5% of the European population severely and permanently diminished by pain and functional impairment. Economically, the current estimated cost of osteoarthritis is 200 billion Euro per year and is set to increase as the European population ages. PurStem results will enable the translation of promising stem cell therapies in this area to the treatment of joint injuries that contribute to the development of osteoarthritis, as well as potential therapies that prevent progression of the disease. These novel regenerative therapies will reduce the economic and social costs of the disease to the European Community.

Problems with repair of bone as a result of trauma, cancer or other causes also exert a heavy economic and social burden on the European community with a potential patient population of over 3 million in the area of bone reconstruction. The primary social benefits from PurStem will accrue from improvements in treatment options in these and other areas which will result in improved quality of life of patients, a reduction in the economic burden associated with chronic disease and cost savings, reproducible treatments and low risk of results dispersion and uncertainties.

Dissemination efforts targeted the general public, within the EU and globally, and patient advocacy groups in the partner countries (Ireland, the UK, Italy and the Czech Republic) in order to achieve maximum awareness of the contribution of PurStem to advances in stem cell therapies and of the potential of stem cells for providing effective therapies in the future. In addition, PurStem’s dissemination strategy specifically targeted school students

Dissemination Activities

Dissemination of the project results to a wide audience, including the general public, students, other stem cell researchers and patient groups, has been achieved using a variety of approaches descried in the following sections.

Project Website

The main dissemination interface of the project with the general public has been the public project website. The website was established early in the project and has been updated regularly with new content and project news. There is a Beginner’s Guide to stem cells for the layperson, as well as links to general stem cell resources. The website includes links to PurStem Facebook and Twitter pages, as well as Wikipedia. In order to reach as large an audience as possible, much of the content on the website is available in the languages of all project partners (English, Czech and Italian).

Seminars

Over the course of the project, a number of PurStem partners have delivered seminars and public presentations. These events have attracted wide audiences, including the general public, undergraduate students, researchers, industry representatives and other EU projects. A selection of the seminars and presentations given is listed here.
• "Stem Cells In 2009 - Less Mystery, More Hopes (Prof. Mokry, CUNI)
• "Stem Cell Therapy: the Stem Cell Host Interaction in Tissue Repair" (Prof. Barry, NUI Galway)
• PurStem was presented at the 4th Annual Stem Cells and Regenerative Medicine World Congress (Jan 2010), South San Francisco, USA (Tomas Soukup, CUNI).
• “MSCs in Vivo” was presented by Dennis McGonagle (UL) at the University of Sheffield, UK, in November 2009.
• PurStem was presented as an example of fruitful academic/industry collaboration at the National Centre for Biomedical Engineering Sciences Industry Outreach Event “Enterprise and Technology – the R&D Challenge” at NUI Galway (Ciaran Clissmann, Pintail).
• EuroStemCell – 22nd June 2010 “NUI Galway Researchers take part in European Stem Cell Public Engagement Project”
• UL Sally Boxall gave a presentation entitled “PurStem experience in mesenchymal stem cell manufacture for therapy: High number or high quality?” at the Leeds Institute of Molecular Medicine, 24th June 2011
• UL Elena Jones gave a presentation entitled “Characterisation of native bone marrow MSCs” and Richard Cuthbert gave a presentation entitled “Clinical-grade selection of bone marrow MSCs” at Miltenyi Biotech Headquarters, Germany, 20th July 2011
• UL Dennis McGonagle gave a presentation at Lonza, Walkersville, USA 8th October 2011
• PurStem project was presented at the prestigious scientific event - Scientia Scientia Pragensis (12 Feb 2011, Prague, Czech Republic).

Engagement with Other Stem Cell Researchers

Dissemination of project results to other stem cell researchers is key to ensuring that the maximum impact is achieved. To this end, PurStem researchers interacted directly with other stem cell researchers locally and engaged with the ISCT:
• The PurStem project was presented to partner CUNI colleagues from the University in Pardubice – two of them joined our team for cell culture training.
• In June 2011, partner UL hosted a visit by Prof. Nzeh, head of the Stem cell facility in Nigeria to discuss the work being done in Leeds, including the PurStem project, to gain a better understanding of the research being carried out in the MSC field and potential applications towards therapy.
• Dennis McGonagle of UL gave a presentation entitled “Towards clinical-scale enrichment of mesenchymal stromal cells for orthopaedic applications” at the 17th ISCT Annual Meeting in Rotterdam, 18th May 2011
• Sally Boxall of UL presented 2 posters entitled “Novel receptor expression during standardised culture expansion of bone marrow multipotential stromal cells” and “High abundance of CD271+ multipotential stromal cells (MSCs) in intermedullary cavity of long-bones” at the 17th ISCT Annual Meeting in Rotterdam, 19th May 2011
• Cindy Coleman of NUIG presented a poster “High Throughput Characterisation of the Mesenchymal Stem Cell Surface Phenotype” at ISCT 2011.

Engagement with Patient Groups

Patient seminars were held where a distillation of the scientific work of the project was presented in addition to material focused on a particular audience or a particular medical condition. The audiences for these seminars included groups of individuals with a given condition (e.g. osteoarthritis), and also more mixed groups including carers and families of affected individuals. The seminar material was created by the scientists who are directly involved in the PurStem laboratories – it combined background information about MSCs (isolation, culturing, etc.) with stem cell work carried out in the partner laboratories (both PurStem and non-PurStem efforts).
• NUIG presented a seminar on “Future Repair Strategies for Osteoarthritis” to the Irish national arthritis patient organisation “Arthritis Ireland” on 7 October 2010 in Dublin. A similar seminar was given to Arthritis Ireland groups in Limerick (31 March 2011) and Tralee (26 February 2011).
• CUNI presented a seminar for “Roska”, a non-profit patient association for people suffering from multiple sclerosis, at Hradec Kralove on 22 November 2010. This focused on the potential value of stem cells as a therapy for multiple sclerosis. Slides in Czech were prepared for the PurStem website.

Schools Outreach

The project’s schools outreach strategy included a dedicated Student Zone on the project website with information in all three project languages and the distribution of school-specific information to secondary schools. Flyers were also created for circulation in schools. The project also took advantage of the dedicated outreach facilities, resources and personnel available within the project partners’ host institutions.
• In July 2009, partner UL hosted a Schools Outreach event in Leeds, UK.
• Partner CUNI hosted a Schools Outreach seminar and practical lab session for local secondary schools in September 2010 in Prague, Czech Republic.

Exploitation

Several strong opportunies have arisen in terms of exploitation within the PurStem project. These reflect the scope of the activity and will lead to direct commercialization efforts, new intellectual property, strengthening of exisitng technology and further lines of investigation. It is anticipated that several significant activities will arise from PurStem and that it will have a substantial future impact. This will apply both to the PurStem and to the wider research community to which the project contributes.

The main avenues of exploitation are in further research and in the commercialisation of the results of the project.

Further Research

The project has enabled the team to build strong networks of contacts and collaboration both within and across its host nations. This has led to new research projects and new opportunities to work with researchers from across the world. Some examples of new academic research which utilises PurStem work include:
- A new study of equine stem cells at NUI Galway. This study has not only added value to the Irish bloodstock industry (an important export engine) but has also contributed to the background research leading to a new FP7 project (REDDSTAR) coordinated by NUI Galway and involving SMEs Orbsen and Pintail.
- New research and pre-clinical trials of stem cells with the Irish Blood Transfusion Services Board, the national body responsible for medical blood supplies and associated products. An important enabler for the new partnership is the availability of the GMP facility at NUI Galway, as well as the GMP experience and expertise developed during PurStem.
- A new diabetic limb ischemia project at REMEDI (NUI Galway) is also leveraging the PurStem results; this project is another important input to the new FP7 REDDSTAR project.
- The University of Leeds (UL) uses expertise from PurStem in a new study with the NHS Blood and Transplant organisation to carry out pre-clinical investigations with MSCs.
- UL is also using PurStem knowledge in an osteoporosis project with the AO Foundation and a joint healing/longevity project with the WELMEC Centre of Excellence.

Of particular value is the new FP7 REDDSTAR project which utilises antigens discovered and developed within PurStem to provide a new approach to stem cell isolation and proliferation. The new technology offers a substantial step forward in terms of specificity, purity and characterisation of stem cells. An ambitious project addressing the use of stem cells in diabetic complications including nephropathy, retinopathy, cardiomyopathy, neuropathy and fracture, and involving leading research teams and SMEs from across Europe and the US, is set to begin in Q3 2012. The project will, like PurStem, be coordinated by REMEDI and involve SMEs Orbsen Therapeutics and Pintail.

Industry-Academic Opportunities

In addition to the academic research which builds on the work of PurStem, several new industry/academic collaborations also leverage these results. While details of these collaborations cannot be provided here for commercial reasons, the following organisations are examples of those with published track records of working with PurStem partners in areas that are relevant to PurStem

Commercialisation Opportunities

A number of the activities of PurStem generated intellectual property which has strong potential for commercialisation and subsequent licensing or other exploitation.

Orbsen Therapeutics has already applied for a patent on the new intellectual property developed during PurStem (United Kingdom Patent Application No. 1202319.8 “Stromal Stem Cells” by Orbsen Therapeutics Limited). This IP is a key asset of the spin-out company and has excellent potential for application in a broad range of therapies. As noted above, it will receive substantial further development in the FP7 REDDSTAR project.

Ovagen has used PurStem to develop and progress its proprietary antibody-generation technology using germ-free chickens. The Ovagen transgenic protein production platform is a novel method for the large scale production of recombinant proteins which is comparable to currently used production platforms with yields in excess of 5mg/ml of recombinant protein. The avian system also provides naturally occurring post-translational modifications that may well be better tolerated in use for human therapeutics in comparison to other production systems such as CHO.
It is anticipated that this technology would be targeted at biotech companies who have an interest in the biosimilars space, particularly for the production of therapeutic recombinant monoclonal antibodies. The technology is ready for use and partnerships can be established with immediate effect and no further research and development is necessary.

The technology as stands is not covered by any means of IPR due to historical reasons where the patents surrounding the technology were allowed to lapse and become public. Unless modifications are made to the production platform there is unlikely to be any IP arising from this foreground.
The potential impact from this technology is a competitive protein production platform that can rival current technologies in yield, cost and technological advantages. As the market for recombinant therapeutic antibodies is in excess of $80bn there is great commercial potential using the Ovagen transgenic platform.

The processes and technologies learnt and applied in PurStem will be invaluable in the further development of the company. PurStem also provided Ovagen with the opportunity to forge and/or reinforce a number of important strategic alliances and links, notably with the Roslin Institute in Edinburgh.
NUIG and CUNI have developed a new, lower-cost recipe for effective serum-free medium for stem cell culture. This medium would be of particular interest to companies and researchers involved in stem cell manufacture and research. Patenting and potential licensing of the medium is under consideration, with priority given to Orbsen Therapeutics.

List of Websites:
Project Website: www.purstem.eu
National University of Ireland, Galway (Coordinator)
Professor Frank Barry
Regenerative Medicine Institute, National Centre for Biomedical Engineering Science, National University of Ireland, Galway, Ireland
frank.barry@nuig.ie

University of Genoa
Professor Ranieri Cancedda
Dipartimento di Oncologia, Biologica e Genetica,
recently changed to Dipartimento di Medicina Sperimentale,
Università di Genova and Instituto Nazionale per la Ricerca sul Cancro,
Genoa 16132, Italy
ranieri.cancedda@unige.it

University of Leeds
Professor Dennis McGonagle
NIHR Leeds Biomedical Medical Research Unit, Chapel Allerton Hospital, Leeds Teaching Hospitals Trust, Leeds, UK LS7 4SA
D.G.McGonagle@leeds.ac.uk

Charles University, Prague
Professor Jaroslav Mokry
Department of Histology and Embryology, Charles University in Prague, Simkova 870, 500 38 Hradec Králové
mokry@lfhk.cuni.cz

Ovagen International Ltd.
Dr. Iain Shaw
Carrentrila, Ballina, Co. Mayo, Ireland.
iain.shaw@ovagen.ie

Orbsen Therapeutics
Dr. Steve Elliman
Orbsen Building, National University of Ireland, Galway, Ireland
steve.elliman@orbsentherapeutics.com

Pintail Ltd.
Mr. Ciaran Clissmann
77 Springhill Ave,
Blackrock, Co. Dublin, Ireland
ciaran.clissmann@pintail.eu

Related information

Contact

Frank Barry, (Scientific Director of REMEDI)
Tel.: +35391495108
E-mail
Record Number: 196524 / Last updated on: 2017-03-27
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top