Evaluation of a cell-based osteogenic formulation compliant with good manufacturing practice for use in tissue engineering

Proper bony tissue regeneration requires mechanical stabilization, an osteogenic biological activity and appropriate scaffolds. The latter two elements can be combined in a hydrogel format for effective delivery, so it can readily adapt to the architecture of the defect. We evaluated a Good Manufacturing Practice-compliant formulation composed of bone marrow-derived mesenchymal stromal cells in combination with bone particles (Ø = 0.25 to 1 µm) and fibrin, which can be readily translated into the clinical setting for the treatment of bone defects, as an alternative to bone tissue autografts. Remarkably, cells survived with unaltered phenotype (CD73+, CD90+, CD105+, CD31−, CD45−) and retained their osteogenic capacity up to 48 h after being combined with hydrogel and bone particles, thus demonstrating the stability of their identity and potency. Moreover, in a subchronic toxicity in vivo study, no toxicity was observed upon subcutaneous administration in athymic mice and signs of osteogenesis and vascularization were detected 2 months after administration. The preclinical data gathered in the present work, in compliance with current quality and regulatory requirements, demonstrated the feasibility of formulating an osteogenic cell-based tissue engineering product with a defined profile including identity, purity and potency (in vitro and in vivo), and the stability of these attributes, which complements the preclinical package required prior to move towards its use of prior to its clinical use.

0.25 to 1 µm) and fibrin, which can be readily translated into the clinical setting for the treatment of bone defects, as an alternative to bone tissue autografts. Remarkably, cells survived with unaltered phenotype (CD73 + , CD90 + , CD105 + , CD31 -, CD45 -) and retained their osteogenic capacity up to 48 h after being combined with hydrogel and bone particles, thus demonstrating the stability of their identity and potency. Moreover, in a subchronic toxicity in vivo study, no toxicity was observed upon subcutaneous administration in athymic mice and signs of osteogenesis and vascularization were detected two months after administration.
Conclusions The preclinical data gathered in the present work, in compliance with current quality and regulatory requirements, demonstrated the feasibility of formulating an osteogenic cell-based tissue engineering product with a defined profile including identity, purity and potency ( in vitro and in vivo ), which is required prior to move towards its use of prior to its clinical use.

Background
Failure of the physiologic reaction to acute or chronic bone disorders (e.g. fractures, nonunions, large defects after trauma or tumors) typically requires surgical intervention and implantation of bone grafts. Autografts and heterologous transplants of bony tissue, as 3 well as the use of implants made of biomaterials, are the most common approaches in today's orthopedics field (1). Among them, the bony autograft sourced from the iliac crest is the preferred treatment option for a wide range of orthopedic conditions, including the management of complex fractures or in non-union defects (1,2). The use of autologous tissue as a vector for bony regeneration fulfils three key requirements: 1) introduces cells with osteogenic potential, 2) offers structural support (osteoconduction), and 3) contributes with growth factors that promote vascularization and osteoinduction (1).
However, the complete substitution of the damaged bony tissue is not always achieved through the use of autografts, which can lead to the failure of the autograft at long term (3). On the other hand, the surgical collection of autologous bone is highly associated with morbidity (4), which in some cases can be overcome by using allogeneic, decellularized human bone from tissue banks. However, such strategy lacks the benefits associated to the regenerating activity displayed by osteogenic cells, which have been demonstrated to be a key factor in our hands, after its successful use in animal models and clinical cases (5,6), in accordance with the "diamond concept" (1). In fact, autografts are preferred for treating large bony defects even when bony tissue from tissue bank is available.
Unfortunately, this option is not valid for all patients and it is not exempt of risks of the procedure required for tissue extraction, as discussed previously. Therefore, when this approach is not feasible (i.e. re-interventions, donor site morbidity, infections), the use of osteogenic cells isolated from bone marrow (BM), either as bulk concentrates or enriched in multipotent mesenchymal stromal cells by ex vivo expansion, is an alternative that has already been explored in large animal models, and also in early Phase I/II clinical trials with encouraging results (5)(6)(7)(8). It is particularly interesting the development of products that combine culture-expanded multipotent Mesenchymal Stromal Cells (MSC) and scaffolds, namely "Tissue Engineering Products" (TEP), resulting in a new medicinal entity 4 with osteogenic potential that is specifically regulated as advanced therapy and needs to comply with pharmaceutical regulations (9).
In the present study, we formulated TEP based on the use of cells with osteogenic potential (namely, ex vivo expanded MSC that were compared to BM concentrates) combined with bony particles from tissue bank embedded in a hydrogel that, altogether, can induce the generation of new tissue while adapting to the diverse architecture of the simulated cylindrical bony defects. In agreement with current quality and regulatory requirements, TEP's stability, osteogenic potential and in vivo safety were assessed comprehensively under Good Laboratory and Manufacturing Practices.

Cell cultures
Clinical grade ex vivo expanded MSC derived from the mononuclear cell (MNC) fraction of BM aspirates were produced in the context of a clinical trial (EudraCT No. 2010-024041-78) with appropriate donor informed consent. Cells were further expanded in vitro up to sufficient numbers (always under passage 4) by using Dulbecco's Modified Eagle's Medium (DMEM; Gibco) containing 2 mM glutamine supplemented with 10% human serum (hSer) B/AB (10,11). All cultures were maintained at 37ºC and 5% CO 2 in humidified incubators.

Stability assessment
The effect of fibrinogen on BM-MSCs' stability was assessed by means of analyzing cell viability, phenotype, capacity to adhere to plastic surfaces and osteogenic differentiation capacity at 2-8 ºC with freshly prepared cellular suspensions (13). Experimental surgery 7 Animals were induced with inhalatory anesthesia consisting of 4% isofluorane mixed with 100% oxygen in an anesthetic chamber. Anaesthetized animals were placed on a heating blanket and maintained on isoflurane 2% mixed with 100% oxygen administered through an anesthetic mask. Two mg/kg meloxicam were administered subcutaneously for analgesia, and ophthalmic lubricant was placed on the eyes in order to prevent eyes dryness. The surgical zone was sterilized and a 1 cm long incision was made in the epidermis of the dorsal zone, caudal to the shoulder blade and perpendicular to the spinal column. In order to perform the incision, the skin was pinched and pulled using bluntended surgical forceps and incised with straight-bladed Mayo scissors. Then, the subcutaneous area was dissected cranially to the incision forming a pouch where the test and reference items were placed using curved Adson forceps. After implantation, the surgical wound was closed with two surgical staples and the animals were maintained on the heating blanket until total recovery before returning them to the original cage (Supplemental Fig. 1).

Necropsy and histology
At the end of the two-month follow-up period, animals were euthanized by an overdose of sodium pentobarbital (200 mg/kg, 60 mg/mL) administered intraperitoneally. The macroscopic analysis included the assessment of the musculature, fur, skin and natural orifices. Additionally, the brain and cranium, thoracic cavity and mediastinum, trachea, esophagus, glands and lymph nodes, lungs, heart, abdominal cavity including stomach, small and large intestine, liver, spleen, kidneys and genitourinary system were examined.
All these organs were then extracted, minced, fixed and embedded in paraffin for further histological studies. Two µm thick microtome sections of the specimens were cut in the sagittal plane and examined by routine hematoxylin and eosin (H&E) staining using a grading score specific for the assessment of bone formation (Supplemental Table 1). 8

Data analysis
Descriptive data was expressed as mean ± standard deviation (number of replicates) or mean (range of values). One-way ANOVA test and Bonferroni's multiple comparison tests were performed to evaluate differences in water and food consumption, body weight gain.
Paired t-tests were performed to evaluate differences in size and histological scores.
Statistical significance was set at * p < 0.05. MSC embedded in fibrinogen survived and maintained their phenotype and osteogenic potential 9 The effect of fibrinogen on cellular stability was investigated by assessing cell viability, which remained stable up to 48 hours at 2-8 ºC (Figure 2A). Phenotype at 24 h was consistent with initial MSC identity determined at the start of the experiment ( Figure 2B).

Sourcing of cells, characterization
In order to understand whether cells retained their capacity to adhere to plastic surfaces along the course of the study, as a surrogate marker of cell viability, we evaluated this parameter at 0, 6, 18, 24 and 48 h to complement cytometric data. Interestingly, cells were viable along the study time and no gross differences were observed regarding the morphology of adherent cells during the first 24 hours ( Figure 2C syringes whose edges were previously cut, with the aim of simulating a cylindrical bone defect (Supplemental Figure 1C). Under such conditions, the jelly mixture clotted within minutes while adapting perfectly its shape to the cylindrical shape.
Next we generated cylindrical constructs as a model experimental situation of cylindrical defect, using three formulations: acellular (control group), MNC-loaded and MSC-loaded TEPs. The cellular doses were 48x10 7 MNC/cm 3 of bone and 60x10 6 MSC/cm 3 of bone ( Figure 3A). Those cylinders were cut in a way that 0.2 cm 3 constructs were obtained for the following in vivo verification of their safety and osteogenic and angiogenic capacity.

In vivo experiments demonstrating vascularization and osteogenesis
In addition to in vitro analyses, in vivo studies were performed to demonstrate safety and efficacy of the osteogenic formulation reported here. Constructs of 0.2 cm 3 were implanted subcutaneously between the scapulae of athymic mice and they were followed up for 2 months. All animals survived the surgery without any adverse reactions to the procedure. Several clinical parameters were monitored throughout the study (including weight, general condition, wound appearance, food and water intake). The surgical wound healed normally in all animals, which presented a healthy general condition throughout the experimental phase. No spontaneous mortality occurred in the course of the study, nor was any pathological condition observed. All animals increased in body weight throughout the study, only a slight decrease in the body weight gain immediately after administration of the constructs was observed, which can be attributed to animal handling and surgery ( Figure 3B). A full necropsy at termination was performed to all animals and only unspecific findings were observed in all three experimental groups, irrespective of treatment, which were not considered to have pathologic relevance thus confirming the lack of toxicity (Supplemental Figure 2). In addition, no signs of rejection of the implants were observed macroscopically.
In the control group, the constructs were absorbed gradually until becoming unnoticeable macroscopically (with the exception of one animal, as shown in Supplemental Figure 3 at week 5). In the groups treated with cell-based TEPs, a reduction of the initial size of the constructs was observed although they increased gradually thereafter reaching a peak of 11 growth at five weeks after surgery and further shrunk until the euthanasia at week 8. At termination, the dimensions of the constructs varied between experimental groups (Ø = 1 ± 2.4, 6.8 ± 6.1 and 12 ±10.0 mm 2 for control, MNC and BM-MSC, respectively), being MSC-loaded TEP the one showing a significantly higher size increase compared to the control group (p = 0.0191) ( Figure 3C). Interestingly, the presence of MSC prevented the fragmentation of the test item that remained in one single piece along the course of the study, whilst multifocal fragmented particles were observed in some animals from the

Discussion
Treatment of bone defects using osteogenic cell-laden hydrogels that can adapt to the architecture of the lesion might soon become a useful tool for orthopedic surgeons (18). In the present study, we explored whether BM-MSC expanded under a Good Manufacturing Practice (GMP)-compliant production process preserved their viability, identity and osteogenic potential when combined with commercial fibrin and decellularized/deantigenized bone particles sourced from tissue bank as a TEP. We found that indeed cells remained viable and preserved their osteogenic capacity in such osteogenic formulation, so it could be potentially used in the treatment of a wide range of orthopedic conditions or even as bioink in 3D bioprinting strategies (19,20). Regarding the phenotype, although HLA-DR expression was higher than the criteria established by the International Society for Cell and Gene Therapy (ISCT) (21), it did comply with product specifications that were approved by the competent authority (11). Indeed, varying percentages of HLA-DR expression are often reported in MSC cultures, despite meeting the rest of defining criteria (22)(23)(24)(25)(26).
In current clinical practice, the use of BM concentrates are becoming popular in an attempt to provide stem cells to the fracture site. Although BM aspirates are typically processed in order to isolate the MNC fraction, this is not composed only of MSC but a heterogeneous population of B-cells, T-cells and monocytes, as well as rare progenitor cell types such as hematopoietic stem progenitor cells (HPCs) and endothelial progenitor cells (EPCs), it is still unclear which component or combination of components exactly determines its tissue induction activity, which can be exerted either by direct differentiation or by paracrine activity. It has been previously described that the MNC fraction from BM can promote angiogenesis (27), mediate vascular repair, produce cytoprotective growth factors and cytokines (28), and regenerate bone (29). From all the different subpopulations present in the bone marrow, MSC contribute to a very small fraction, estimated in the range from 0.001-0.1% of MNC (30,31). Despite of such low 13 occurrence, BM-MSC can be efficiently expanded ex vivo and induced to differentiate into multiple lineages when subjected to defined culture conditions (29). In the orthopedics field, the use of pure populations of MSC is thought to promote bone formation more efficiently.
In the present study we employed fibrin hydrogel as a clotting agent, which is a commercially available product for clinical use that can also be manufactured in situ on demand either as allogeneic or autologous product (32,33). Typically, fibrin is presented in a formulation of two components: A) concentrate or purified fibrinogen, which is the precursor glycoprotein of fibrin, and B) a mixture of factor XIII, thrombin and calcium, which triggers the polymerization reaction (32). In fact, fibrin glue is a widely known product in the surgery field that is used for rapid hemostasis (34) sheets are gaining ground among the regenerative medicine field suggesting that different delivery modes of the active ingredient within the TEP may enhance tissue regeneration and therefore much effort is currently being made in order to improve such formulations (44,45).

Conclusions
The present study provides evidence of the feasibility of TEP preparation with clinicalgrade reagents while preserving the identity, osteogenic potency and safety of cells used in its formulation. Moreover, the preclinical data gathered in the present work, in