BASIC AND TRANSLATIONAL ASPECTS OF CELL-BASED APPROACHES FOR EARLY AND LATE STAGE OSTEOARTHRITIS

Osteoarthritis (OA) is a highly disabling pathology which is worldwide investigated by the scientific community due to its increasing diffusion. The incidence of this age-related disease is increasing with population ageing and worldwide estimates indicate that 9.6% of men and 18% of women with more than 60 years present OA-related symptoms. Osteoarthritis induces the progressive damage of articular cartilage and subchondral bone and can eventually lead to the complete loss of joint functionality. Nowadays, the management of OA includes non-pharmacological, pharmacological, and surgical treatments. Non pharmacological approaches include exercise, weight loss, and physiotherapy and are used in conjunction with pharmacological treatments. The pharmacological management of OA patients is mainly based on the use of anti-inflammatory drugs for pain control. Hence, the pharmacological approach to OA patients is synptomatic, but not resolutive, since it is not able to alter the progression of the disease. These therapeutical options are not suitable for late stage OA patients, whereby the severe pain and the functional limitations caused by advanced OA degeneration are currently resolved by joint replacement. Due to the low self-repair ability of articular cartilage, untreated cartilage lesions can easily degenerate into OA. Different strategies have been developed to treat promptly chondral lesions, comprising cell-based therapies, such as autologous chondrocyte implantation (ACI). These cell-based therapies have been recently proposed also for the treatment of early OA patients in order to overcome or at least delay the need for a more invasive intervention such as joint replacement. However, major issues associated to the clinical use of autologous articular chondrocytes (ACs) are the limited number of cell harvestable from a small cartilage biopsy and the de-differentiation process occurring during the expansion phase required to achieve a clinically relevant number of cells. Furthermore, advanced age and pathological state of joints negatively affect the chondrogenic potential of ACs, representing important factors to be considered when applying chondrocyte-based therapies in particular categories of patients. In view of a possible use of autologous cell-based therapies for OA patients, we analyzed specific features of ACs as cellular yield, cell doubling rates and the dependence between these parameters and patient-related data in a set of 211 OA patients undergoing total joint replacement (Chapter 3). The patient age was not statistically correlated to the cellular yield, but was negatively correlated with the proliferation rate. No significant correlation was observed between the level of cartilage degeneration (ICRS score) and cellular yield and proliferation rates. However, in samples with the highest degree of cartilage degeneration (ICRS score 4) the cellular yield was lower compared to the other three groups (ICRS scores 1-3). In conclusion, we found that age and degenerative state of cartilage affect some basic parameter of articular chondrocytes and should be considered when evaluating the quality of the cell source to be used in clinical applications. To overcome some of the limitations related to the use of ACs, mesenchymal stem cells (MSCs) present in several tissues, such as bone marrow and fat, have been proposed as cell candidate for cartilage treatment thanks to their demonstrated chondrogenic ability. When developing treatments for knee chondral lesions, infrapatellar fat pad and knee subcutaneous adipose tissue, two fat depots easily accessible during knee surgery, can be considered appealing sources for MSCs harvesting. We performed a donor-matched comparison between infrapatellar fat pad MSCs (IFP-MSCs) and knee subcutaneous adipose tissue stem cells (ASCs) obtained from OA patients undergoing total knee replacement, analyzing their immunophenotype and multi-differentiative ability, with a focus on their osteogenic and chondrogenic potential (Chapter 4). We found that these cell populations share common features at the undifferentiated state, such as immunophenotype and clonogenic potential, but display a specific commitment towards the osteogenic and chondrogenic lineage. Indeed, significantly higher levels of osteogenic markers, as calcified matrix deposition and alkaline phosphatase activity, were found in ASCs, highlighting the superior osteogenic commitment of this cell population in comparison with IFP-MSCs. Conversely, IFP-MSCs differentiated towards the chondrogenic lineage by 3D pellet culture showed greater glycosaminoglycans (GAGs) deposition and higher expression of chondrogenic genes as aggrecan (ACAN) and type II collagen (COL2A1) compared to ASCs pellets, revealing a superior chondrogenic potential. This result was supported by lower expression of hyperthrophic and fibrotic markers, as type X (COL10A1) and type I (COL1A1) collagen, in IFP-MSCs pellets compared to ASCs. The observed dissimilarities indicate that populations of adipose-derived MSCs harvested from different anatomical sites have specific features that suggest their preferential use for specific cell-based applications. In the last decade, the co-culture of ACs and MSCs has gained growing interest to investigate the cross-talk between these cell types and to evaluate the possibility of replacing a fraction of ACs with MSCs, in order to reduce the in vitro expansion phase of ACs and subsequently limit their de-differentiation. Controversial results have been reported on the possibility to generate cartilage-like matrix by co-culturing ACs and MSCs, indicating a prominent role for the origin and de-differentiation state of chondrocytes as determinants of the outcomes. We evaluated whether the co-culture of IFP-MSCs and ASCs with a small fraction of ACs was able to lead to cartilage-like matrix generation and to the expression of chondrogenic markers comparable to ACs (Chapter 5). To better resemble a possible clinical application of this strategy we performed autologous co-cultures combining IFP-MSCs and ASCs with expanded and cryo-preserved donor-matched ACs and we used cells derived from OA patients so as not to neglect the impact of the pathological and de-differentiated state of ACs on the outcome of the co-cultures. Chondrogenic genes SRY (sex determining region Y)-box 9 (SOX9), COL2A1, and ACAN were less expressed in co-cultures compared to ACs mono-cultures. No significant differences were observed for GAGs/DNA in mono-cultures, demonstrating the reduced chondrogenic potential of ACs, probably deriving by both their pathological origin and their de-differentiated state. Total GAGs content in co-cultures did not differ significantly from values predicted as the sum of each cell type contribution corrected for the co-culture ratio, as confirmed by histology. Therefore, a small percentage of expanded and cryopreserved ACs did not lead to IFP-MSCs and ASCs chondro-induction. Our results suggest that chondrogenic potential and origin of chondrocytes play a relevant role in the outcome of co-cultures, indicating the need for further investigations to demonstrate their clinical relevance in the treatment of aged osteoarthritic patients. As aforementioned, chondrogenic differentiation of both ACs and MSCs is usually performed in 3D culture models that partially mimic the 3D environment of cartilagineous matrix such as culture in pellets and in 3D scaffolds. Pellet culture is widely used in the field of cartilage tissue engineering to test and optimize differentiation protocols. These experiments normally require to prepare replicates from a relevant number of donors to investigate different culture conditions. Unfortunately, handling a high number of samples in 3D pellet culture is extremely time consuming and is a major issue in experiments with many different conditions. To overcome these limitations, we exploited low-cost rapid prototyping techniques, such as laser ablation and replica molding, to generate multi-well chips in polydimethylsiloxane (PDMS) for the formation and culture of cell aggregates (Chapter 6). Each PDMS chip allowed the simultaneous culture of several pellets leading to a significant reduction in the time required for pellet seeding and medium refresh operations. Proliferation and metabolic activity were comparable between pellets of ACs cultured in PDMS chips and pellets cultured in polypropylene tubes, whereas type II collagen deposition was increased in pellets cultured in the PDMS chips. The same rapid prototyping approach was applied to generate a patterned scaffold. As a proof of concept, we biofabricated a multi-well implantable constructs using clinically approved fibrin glue. We demonstrated that regions with different cell densities can be generated within the construct and that cells can be distributed in a spatially controlled way. These multi-well implantable scaffolds can be seeded with multiple cell types allowing the precise distribution of different cell populations, thus representing a useful tool to investigate in vitro and in vivo cell interactions in a 3D environment. The introduction of engineered tissue into the clinical practice implies the achievement of high quality and safety standards. The use of automated bioreactor-based manufacturing can significantly increase the reproducibility of the results, being minimally affected by operator variability. Furthermore dynamic culture systems, such as perfusion bioreactors, represent a smart approach to overcome the limitations encountered in the static culture of 3D constructs having clinically relevant dimensions. Indeed, perfusion bioreactors have been shown to increase the homogeneity of cell distribution within the scaffold compared to static cell seeding. Perfusion culture provides a continuous and homogeneous influx of fresh medium throughout the construct, improving the quality and homogeneity of the extracellular matrix. The use of optimized protocols for the expansion and re-differentiation of ACs employing clinically approved growth factors is another key step to achieve a successful outcome and to grant the clinical translability of the results. Thus, in collaboration with the group supervised by Prof. Frédéric Mallein-Gerin (Institut de Biologie et Chimie des Proteines, Lyon, France) we tested the combination of a specific cocktail of clinically approved growth factors with a bi-directional perfusion bioreactor (OPB, Oscillating Perfusion Bioreactor) with the aim of improving cartilage matrix production (Chapter 7). We established a perfusion program including phases of high and low perfusion speeds to alternate sequences of cell stimulation and matrix deposition and we compared collagen sponges seeded and cultured in dynamic conditions with sponges seeded and cultured in standard static conditions. We found that perfusion improved cartilage matrix deposition within the sponges, in comparison with static conditions. More precisely, in the sponges cultured in the bioreactor a cartilaginous matrix rich in type II and type IX collagen and GAGs, with no signs of hypertrophy, was produced. Furthermore, a lower amount of type I collagen was produced in the sponges cultured in dynamic conditions, indicating that perfusion limits the risk of fibrocartilage formation. In conclusions, the dynamic culture by means of a perfusion bioreactor combined with the use of a cocktail of clinically approved growth factors proved to be effective for the generation of engineered cartilagineous grafts, improving the homogeneity and quality of the extracellular matrix generated by articular chondrocytes seeded within collagen sponges. Aged patients with late stage OA involving an extensive erosion of articular cartilage and exposure of the underlying subchondral bone cannot be treated with cell-based therapies or standard surgical approaches aiming at cartilage restoration. In these patients, joint replacement is currently the only clinical option to restore the articular function. The insufficient implant stability is an important determinant in the failure of cementless prostheses, leading to an increase in the risk of implant mobilization and in the number of revision procedures, with major drawbacks for patients and for National Health Systems. With the aim of increasing implant osseointegration, porous metallic materials have been developed and applied to cementless implant technology to maximize bone ingrowth and bone-implant contact. More recently, the enrichment of porous titanium with growth factors has been proposed as a novel strategy to further improve implant osseointegration. We combined a macroporous titanium (Trabecular Titanium™, TT), currently used in the clinical practice, with a biocompatible hydrogel (amidated carboxymethylcellulose, CMCA) able to encapsulate osteoinductive factors and osteoprogenitor cells (Chapter 8). In particular, we chose strontium as osteoinductive factor and bone-marrow derived MSCs (BMSCs) as osteogenic progenitor cells for implant enrichment. We tested different concentrations of SrCl2 to select the optimal concentration to induce BMSCs differentiation. We found that SrCl2 at 5 µg/ml significantly increased calcified matrix deposition, type I collagen (COL1A1) expression and alkaline phosphatase activity, being the most effective concentration among the ones tested. The enrichment of TT with CMCA (TT+CMCA) significantly increased cell retention within the implant, representing an important improvement in view of a future clinical application of the bioactive implant. Furthermore, we found that BMSCs cultured in TT+CMCA in the presence of SrCl2 underwent a more efficient osteogenic differentiation, as demonstrated by higher alkaline phosphatase activity and calcium levels. Based on these in vitro results, we performed an in vivo study to evaluate the performance of the bioactive implant generated combining the macroporous titanium with strontium-enriched CMCA and BMSCs (Chapter 9). To mimic implant-bone interaction, an ectopic model was developed grafting TT implants into decellularized bone seeded with human BMSCs. TT was loaded or not with strontium-enriched CMCA and/or BMSCs and constructs were implanted subcutaneously in athymic mice. Osteodeposition at the bone-implant interface was investigated with fluorescence imaging, micro-CT, scanning electron microscopy (SEM), histology and biomechanical testing at different time points. Micro-CT analysis demonstrated the homogeneity of the engineered bone in all groups, supporting the reproducibility of the ectopic model. Fluorescence imaging, histology, SEM and pull-out mechanical testing showed superior tissue ingrowth in TT implants loaded with both strontium-enriched CMCA and BMSCs. In our model, the synergic action of the bioactive hydrogel and BMSCs increased both bone deposition and TT integration, indicating that the generation of bioactive implants able to promote bone deposition is a promising strategy for the improvement of implant osseointegration. Joint inflammation is an important trait of osteoarthritis, indeed synovitis is observed in various degrees in OA patients, with OA synovium showing a mixed inflammatory infiltrate mainly consisting of macrophages. The central role of macrophages in OA evolution supports the interest in investigating their response to signals typical of OA joints, such as the ones contained in the synovial fluid (SF) of OA patients. In collaboration with the Connective Tissue Cells and Repair group supervised by Prof. Gerjo J.V.M. van Osch (Erasmus MC, Rotterdam, The Netherlands), we investigated the effect of SF from OA patients on the transcriptional expression of anti-inflammatory and pro-inflammatory mediators in monocyte-derived macrophages (Chapter 10). SF from donors without any joint pathology was used as control, and the effect of the different types of SF was tested on non-activated macrophages (M0) and on macrophages activated by IFNγ and TNFα (M1 activation) or IL4 (M2 activation). We found that IL10 and IL1ra were modulated by the treatment with OA SF in M0 and M2 conditions. The effect on IL10 was lost in M1 conditions probably due to the high up-regulation induced by the treatment with IFNγ and TNFα. Our results support the idea that arthritic SF is for macrophages a less pro-inflammatory environment compared to control SF. This may mimic the macrophage situation in the joint, where feedback mechanisms are induced to counteract the ongoing pro-inflammatory processes. Altogether our findings highlight the complexity and multiplicity of factors that should be considered for the development of successful cell based treatments for early and late OA patients. We found that basic features of articular chondrocytes were affected by donors age and cartilage degeneration. We also found that expanded OA ACs did not exert a significant chondroinductive effect on co-cultured MSCs, suggesting that the evaluation of the clinical relevance of this approach in OA patients should strongly consider the impact of the de-differentiation state and pathological origin of ACs. Further investigations in this field are needed to verify whether it is possible to replicate the results obtained using non-expanded healthy ACs to co-cultures using expanded ACs from aged patients affected by osteoarthritis. These investigations may require the optimization of clinically approved protocols for expansion and re-differentiation of OA chondrocytes as well as the implementation of three-dimensional and dynamic culture systems. We demonstrated that low-cost rapid prototyping techniques are a useful tool to develop high-throughput systems for the culture of 3D cell aggregates and to generate patterned scaffolds that allow the controlled spatial disposition of different cell types. We also showed that the use of perfusion bioreactors for the dynamic culture of 3D constructs can improve the quality of the engineered tissue, granting a more homogeneous and efficient distribution of nutrients and cells throughout the construct. In order to overcome the limitations related to the use of ACs, MSCs can be considered as promising alternative cell candidates. Here, the identification of specific subsets of MSCs displaying a peculiar commitment should drive the selection of the most suitable cell source. Indeed, as we demonstrated for IFP-MSCs and ASCs, despite common features, specific populations of MSCs harvested from different anatomical sites have intrinsic features that make them more or less suitable for a specific application. Cell-based therapies aiming at cartilage restoration cannot be currently applied to late stage OA patients. However, in this category of patients cell-based approaches can be developed to face issues that are critical in implant technology, such as poor osseointegration. As we proposed here, osteoprogenitor cells, osteoinductive factors, and macroporous metals can be combined to generate a bioactive implant with improved osseointegration ability. This approach could be used to generate “off-the-shelf” implants pre-loaded with osteoinductive factors that can be seeded with autologous BMSCs with a fully intra-operative approach. Finally, when developing cell-based approaches for OA patients, the inflammatory state of the OA joint cannot be neglected. We found that macrophage transcriptional expression was affected by OA synovial fluid. This suggests that OA environments may also induce alterations in the re-implanted cells and highlights the need for better in vitro models to evaluate the response of cells and engineered tissues to a diseased environment. In conclusion, to develop novel cell-based approaches for the treatment of OA basic and translational studies should take into account multiple factors that may affect the clinical outcome, such as the impaired phenotype of OA chondrocytes, the peculiar commitment of specific populations of MSCs, and the impact of inflammatory environment of OA joints.