Characterization of Platelet Lysate Cultured Mesenchymal Stromal Cells and Their Potential Use in Tissue-Engineered Osteogenic Devices for the Treatment of Bone Defects

Introduction

Mesenchymal stromal cells (MSCs) are nonhematopoietic multipotent stromal cells,^1–3 defined by a combination of physical, morphologic, phenotypical, and functional properties.^4,5 Bone marrow (BM) is considered the most common source of MSCs, but they can be also isolated from other tissues.^6–8 The multipotent capacity of these cells, the possibility of being easily isolated from BM aspirates, and their high ex vivo expansion potential^9,10 make these cells among the best candidates for cell therapy to regenerate injured tissues, particularly for bone diseases. Under specific stimulation, in fact, MSCs undergo osteogenic differentiation through a well-defined pathway.^11,12

One of the great properties of bone is its ability to heal without scarring. However, perturbations of fracture sites may disrupt the repair process, and evidence has been provided that the number and the functional properties of MSCs are reduced in severe bone fractures complicated by nonunions.^13,14

Various types of tissue engineering techniques, ¹⁵ such as the implantation of bone substitute, ¹⁶ the local delivery of growth factors (GFs),^17,18 the injection of bone progenitor cells, ¹⁹ or the stimulation of the progenitor cells pool ²⁰ have been envisaged to facilitate bone healing.

Cell-based tissue engineering for bone regeneration by autologous MSC transplantation combined with a three-dimensional (3D) scaffold may represent a promising therapeutic approach. Several scaffolds are currently available, of both natural and synthetic origin. In vitro–expanded MSCs seeded onto a hydroxyapatite (HA)/tri-calcium phosphate biomaterial have been used in several experimental animal models,^21–25 showing a great potential in promoting bone union. In addition, formation of new vessels is also necessary for osteogenesis, ²⁶ and both processes (new bone deposition and revascularization) are equally important. A variety of molecules are involved in neo-angiogenesis—in particular, the vascular endothelial growth factor (VEGF).^27,28 Several researchers have focused their attention on a combined use of MSCs (loaded or not into a scaffold) and platelet-rich plasma (PRP) or platelet-gel ²⁹ for their significant GFs release,^30–32 including VEGF, ³³ platelet-derived growth factor (PDGF), transforming growth factor beta (TGF-β), insulin-like growth factors, epidermal growth factor (EGF), and basic fibroblast growth factor (bFGF), all representing powerful angiogenic inducers, and mitogenic stimuli for MSCs. ³⁴ To date, in humans, only few clinical trials have been published using ex vivo–expanded MSCs with porous scaffolds for the repair of critical-sized long bone defects.^35–38

Clinical applications of cell therapy represent a complicated challenge, due to the necessity of operating under strict Good Manufacturing Practice (GMP) conditions, ³⁹ to ultimately certify the quality of the cell product. ⁴⁰ This quality assurance includes not only the determination of the sterility of the cell lot and the compliance with specific clinical needs, but also the identity, purity, safety, and functionality of the MSC-based constructs. To culture MSCs in a clinical-grade medium facilitating the translation into clinical practice, platelet lysate (PL) has been used to replace fetal bovine serum (FBS).^41–45 PL medium, besides avoiding the use of animal-derived material, accelerates MSC expansion, reducing the duration of ex vivo manipulation. Another relevant aspect is the possibility that MSCs may undergo malignant transformation, as reported both in humans ⁴⁶ and in mouse models, ⁴⁷ with contradictory results. ⁴⁸ Previous publications have shown that c-myc protein overexpression is linked to transformation in human MSCs. ⁴⁶ On the contrary, absence or decreased p53 expression is well known to be relevant for the process of neoplastic transformation of various cell types. ⁴⁹

Here we report a validation of a manufacturing procedure of a tissue-engineered cell product, composed of an HA scaffold (Orthoss^®) combined with BM-derived MSCs, comparing FBS versus PL as MSC growth media. PL-cultured MSCs retain their typical characteristics and can adhere, uniformly distribute, proliferate, and differentiate inside an HA-based clinical-grade scaffold. Moreover, manipulated cells do not show any propensity to malignant transformation. These translational data bring from bench to bedside an innovative technology for the treatment of bone defects.

Materials and Methods

Results

Analysis of MSC growth rate and transformation propensity

First, we have compared the isolation and culture of MSCs (n = 6) using either FBS or PL medium. Figure 1 shows MSC cultures at p3, grown in FBS or PL medium (Fig. 1a), CFU-F histogram (Fig. 1b), and PD diagram (Fig. 1c). BM cells generated CFU-F at a higher degree in PL medium (n = 3) than in FBS medium (Fig. 1b, p-value = 0.04). The mean of CFU-F per 7.5 × 10⁵ nucleated cells was significantly higher in PL cultures (14.25; range 7.75–21.38) than in FBS cultures (10.22, range 4.92–17.75).

OPEN IN VIEWER

FIG. 1.

Culture of mesenchymal stromal cells (MSCs) under Good Manufacturing Practice conditions in fetal bovine serum (FBS) versus platelet lysate (PL) medium. MSCs in both media show the typical fibroblastic morphology (a, FBS and PL; scale bar = 100 μm). (b) Colony-forming unit–fibroblast (CFU-F) assay: FBS- and PL-MSC clonogenicity. Student's t-test, *p-value = 0.04. The bars represent averages and standard deviations (SDs) of three different donors. (c) Evaluation of population doubling (PD) on FBS-MSCs (continuous line) and PL-MSCs (dashed line) over 16 days. Student's t-test, *p-value < 0.05. The bars represent averages and SDs of three different donors.

At each passage (until p3) cells produced a subconfluent layer (70–80% confluence) with a fibroblastic shape. The cells showed a more elongated morphology (Fig. 1a) when supplemented with PL medium, as already described in the literature. ⁴² PL-cultured MSCs proliferated faster than FBS-cultured MSCs, as shown by the PD analysis, resulting in a significant difference in the number of PD (Fig. 1c; the difference acquired a statistically significant value [p < 0.05] at time point 6 till time point 16). Cell viability at p3 (as measured by Trypan blue exclusion) was always, in both conditions, greater than 90%.

To analyze differences in the release of GFs possibly responsible of accelerated growth rate,^56–58 ELISA tests were performed to determine the amounts of six human GFs. The 12 PL lots used for MSC cultures were assayed for PDGF-αα, PDGF-ββ, TGF-β1, VEGF, EGF, and bFGF contents. All PL lots contained about 10⁹ platelets/μL. The minimum and maximum values, the average, and the SDs of 12 lots are shown in Table 1. The GF levels analyzed in human plasma were undetectable (data not shown). Analysis on FBS always produced undetectable levels, as expected for a bovine-derived product (data not shown).

OPEN IN VIEWER

Table 1.

Quantification of Growth Factor Release in the 12 Platelet Lysate Lots Utilized for Mesenchymal Stromal Cell Expansions

	Min. (pg/mL)	Max. (pg/mL)	Average ± SDs
PDGF-αα	157.68	662.73	312.24 ± 140.65
PDGF-ββ	127.03	449.69	239.72 ± 104.63
TGF-β1	451.91	1998.1	895.96 ± 473.13
VEGF	0.00	476.23	172.97 ± 149.59
EGF	20.743	64.831	33.56 ± 13.83
FGF	13.46	37.10	26.70 ± 6.46

The table shows minimum and maximum levels, averages, and SDs for each growth factor tested.

VEGF, vascular endothelial growth factor; FGF, fibroblast growth factor; PDGF, platelet-derived growth factor; TGF-β, transforming growth factor beta; EGF, epidermal growth factor; SD, standard deviation.

Since accelerated growth rate may be responsible for augmented propensity for spontaneous cell transformation, we performed at p3 three different assays (Fig. 2). Classic cytogenetic analysis was performed with conventional Q-bands by fluorescence using quinacrine (QFQ) banding (Fig. 2.1). All donors were characterized by a normal karyotype, both in FBS (Fig. 2.1a) and PL medium (Fig. 2.1b). Moreover, MSCs were seeded for growth in methylcellulose, and no colony was detected in all MSC cultures, analyzed after 4 weeks (data not shown).

OPEN IN VIEWER

FIG. 2.

Analysis of MSC transformation propensity. Figure 2.1: Cytogenetic analysis of MSCs cultured with FBS (a) or PL (b) medium. A representative cell expansion (out of three performed for both conditions) is shown. Figure 2.2: quantitative reverse transcriptase–polymerase chain reaction (qRT-PCR). Analysis of c-myc (a) and p53 (b) in FBS- or PL-cultured MSCs at passage 3. Peripheral blood mononucleated cells (PBMCs) were utilized as calibrate sample (gray bar) and HL-60 as positive (a) or negative (b) control. Averages and SDs of three experiments performed are shown, Student's t-test, *p-value < 0.05. 2^-DDCt = the qRT-PCR data are reported as a difference between DCt of the target gene (FBS- or PL-cultured cells) and PBMC.

As far as concerns p53 and c-myc gene expression analysis, the mean values and the SDs measured at p3 (n = 3) are shown in Figure 2.2. We found a 4.25 mean fold increase of c-myc expression in HL-60 cell line compared to a 0.85 (FBS-cultured MSCs) and 0.65 (PL-cultured MSCs), with a p-value <0.05 for both conditions (Fig. 2.2a), when compared to HL-60. On the contrary, p53 (Fig. 2.2b) was expressed by MSCs in both conditions (FBS or PL), even though at different levels, but it was completely absent in HL-60 cell line, as expected.

MSC phenotypical and functional characterization

PL- and FBS-cultured MSCs were analyzed for their phenotype at p3 (Fig. 3) and for their multilineage differentiation capacity (Fig. 4). PL-MSCs were constantly negative for typical hematopoietic markers and HLA-DR. On the contrary, pL-MSCs showed a high expression of CD73, CD90, CD105, and HLA-ABC (Fig. 3). Similar levels were observed for FBS condition. The table in Fig. 3 shows the complete panel (n = 6) of averages of percentages and SDs for all tested antigens for both conditions.

OPEN IN VIEWER

FIG. 3.

Immunophenotype of MSCs cultured with FBS or PL medium. A representative cell expansion (out of six performed for each medium) is shown. The antigens CD105, CD90, CD73, human leucocyte antigen clone I (HLA-ABC), CD33, CD34, CD14, CD19, and human leucocyte antigen clone II (HLA-DR) were analyzed by a flow cytometry analysis software (FACS). The histograms show the percentages for each single marker for the selected quadrants, after gating on viable CD45-negative cells. The table represents the averages ± SDs of the nine antigens (n = 6) analyzed for both FBS and PL conditions.

OPEN IN VIEWER

FIG. 4.

Differentiation into mesodermal lineages of FBS-MSCs (a, c, e) or PL-MSCs (b, d, f ). (a, b) Adipogenic differentiation, oil red O staining (20 × ). (c, d) Chondrogenic differentiation, safranin O staining (4 ×). (e, f ) Osteogenic differentiation, alizarin red S staining. A representative cell expansion (out of three performed for each medium) is shown.

MSCs were then induced to differentiate into adipogenic, chondrogenic, and osteogenic lineages and analyzed for their differentiation capacity (Fig. 4). Results demonstrated that MSCs cultured in PL medium (Fig. 4b, d, f) were able to differentiate into adipocytes (Fig. 4b), chondrocytes (Fig. 4d), and osteoblasts (Fig. 4f), similarly to FBS-cultured MSCs (Fig. 4a, c, e, respectively). With respect to PL cultures, we observed a more robust and earlier osteogenic differentiation than in FBS-ones, as described in details in the following paragraph.

MSC osteogenic differentiation: Mineral matrix deposition and gene expression analysis

Both PL- and FBS-MSCs were cultured in osteogenic medium, and at 4, 8, 11, 14, 18, and 21 days, gene expression analysis (n = 2) and alizarin red staining (n = 6) were performed (Fig. 5). When exposed to osteogenic medium over 21 days, PL-cultured MSCs underwent more robust and earlier osteogenic differentiation (only 8 days in PL cultures compared to 14 days in FBS cultures, in the case of the experiment shown in Fig. 5.2), generating more extracellular mineralization than FBS-cultured MSCs, as shown by alizarin red S staining.

OPEN IN VIEWER

FIG. 5.

MSC osteogenic differentiation on monolayer culture. FBS- and PL-MSCs were cultured in osteogenic induction medium at 4, 8, 11, 14, 18, and 21 days. Figure 5.1 shows mRNA expression levels of alkaline phosphatase (ALP), collagen type1 (COL1), osteonectin (ON), runt-related transcription factor 2 (RUNX2), osteopontin (OP), and osteocalcin (OC) on differentiated cells under FBS (continuous line) and PL (dashed line). Averages and SDs of two independent experiments are shown. Figure 5.2 shows alizarin red S staining of mineralized matrix of cells cultured under the indicated conditions. The shown experiment is one of the two analyzed by qRT-PCR in Figure 5.1.

The osteogenic differentiation was confirmed by qRT-PCR, under osteogenic permissive condition, during a 21-day in vitro time-course study. The expression of the six typical osteogenic genes has been observed both in FBS and PL cultures. Interestingly, the expression of all the factors ALP, COL1, ON, RUNX2, OP and OC is significantly higher in PL-cultured cells. Of note, in PL medium the analysis was stopped at 11 days because of the inadequate quality of RNA likely due to the low number of viable cells observed and to an abundant quantity of mineral matrix.

MSC culture onto HA scaffolds

Further experiments were performed only using PL-cultured MSCs, to manufacture a GMP-grade construct. Cells were seeded onto the Orthoss scaffolds and cultured in PL medium. Several culture conditions and quality controls were assessed.

First of all, a clinical-grade precoating material (retronectin) was compared with standard fibronectin, for supporting MSC adhesion on HA. After 2 weeks of culture, we could observe a detectable cell adhesion in both conditions (Fig. 6.1), with no differences between fibronectin-precoated (Fig. 6.1a) and retronectin-precoated (Fig. 6.1b) scaffolds. Following experiments were performed only with retronectin-precoated scaffolds.

OPEN IN VIEWER

FIG. 6.

Scanning electron microscopy analysis. Figure 6.1: PL-cultured MSC adhesion on the fibronectin-precoated (a) or retronectin-precoated (b) scaffold. MSCs were cultured for 2 weeks in PL medium. Figure 6.2: MSC distribution onto the internal (a) and external (b) surfaces. Scaffolds were fractured and processed for SEM analysis after 2 weeks of culture with MSCs in PL medium. Asterisks show broken surfaces without cells, while arrows show the external surface with adherent cells.

Cell distribution on the 3D matrix was therefore determined, to observe if our method could facilitate the flow of the cells into the most internal pores of the scaffolds (Fig. 6.2). Scaffolds cultured for 2 weeks with MSCs were processed for SEM analysis and fractured in the middle. Figure 6.2 demonstrates an equal cell distribution both in the internal (white asterisks, Fig. 6.2a) and in the external (white arrows, Fig. 6.2b) surface. To observe more optical plans of the scaffold and confirm the previous results, immunofluorescence experiments were then performed (Fig. 7). Serial sections of the scaffold, cultured with MSCs for 2 weeks, have been treated by specific staining to observe cytoplasm (with phalloidine, which stains for actin cytoskeleton filaments) and nuclei (with TOTO3). Figure 7 demonstrates adherent cells along the entire perimeter of a scaffold hole.

OPEN IN VIEWER

FIG. 7.

Confocal microscopy image of a scaffold cryosection. Cells were immunolocalized with phalloidin (red), and nuclei were stained with TOTO3 (blue). The boxes show high magnification of a selected area.

Subsequently, we have evaluated the viability of PL-cultured MSCs inside the scaffold and their ability to proliferate among the HA meshes and to colonize the entire area (Fig. 8). At 4 weeks of culture we could observe several layers of MSCs adherent to the scaffold (Fig. 8.1b), demonstrating MSC proliferation inside the scaffold. An identical monolayer was not observed everywhere along the scaffold, and it is likely that variability in cell concentration is maintained (Fig. 8.1c, single cells vs. 8.1 b confluent layer). Cell proliferation was further demonstrated by counting cells on 10 × fluorescence sections images elaborated with the software Adobe Photoshop CS3, as shown in Figure 8.2 (n = 2).

OPEN IN VIEWER

FIG. 8.

Cell proliferation and differentiation on the scaffold. Figure 8.1: SEM analysis of seeded MSC onto the scaffold at 4 weeks. Cells form confluent layers (c). Variability in cell concentration has been always observed. Single cells (c) versus confluent layer (b) are shown. (a) Control scaffold without cells. Figure 8.2: Estimation of total cell numbers on the scaffold on days 4, 14, and 28. A representative experiment of 2 is shown. Ten serial sections were analyzed for each scaffold at each time point, and data show averages and SDs of the corresponding 10 counts of the analyzed sections. Similar values were observed for the second donor analyzed (data not shown). Figure 8.3 shows mRNA expression levels of ALP, COL1, ON, RUNX2, OP, and OC on differentiated PL-MSCs. Averages and SDs of two independent experiments are shown.

Finally, we have demonstrated that PL-cultured MSCs were able to differentiate inside the scaffold into osteogenic lineage, even if slower than in a monolayer culture on plastic dishes (Fig. 8.3). In particular, at 14 and 21 days, increased expression of all the evaluated genes was detected. OP mRNA was upregulated from the very beginning of the culture. Figure 8.3 shows the transcription profile of ALP, COL1, ON, RUNX2, OP and OC of PL-cultured MSCs inside the scaffold (n = 2).

Discussion

Our article describes a complete GMP-grade validation of a process for the manufacturing of a tissue-engineered cell product, composed of BM-derived PL-cultured MSCs combined with an HA scaffold (Orthoss^®) to be used for the treatment of bone defects.

Because of the multiple limitations associated with the use of bone autograft ⁵⁹ or allograft, ⁶⁰ tissue engineering has gained renewed interest. Recent studies have investigated the biological causes of bone degenerative processes (such as nonunions), providing evidence that the number, the proliferation capacity, the physical properties, and the repair ability of osteoblasts are reduced at sites proximal to nonunions.^13,14 For these reasons, novel cell–based tissue-engineered approaches, which can overcome these cell abnormalities, are strongly warranted. To date, in humans, only few clinical trials have been published using ex vivo–expanded MSCs in conjunction with porous scaffolds for the repair of critical-sized long bone defects.^35–38 This difficulty is definitely due to the necessity of operating under GMP conditions, ³⁹ which also implies the strict definition of parameters allowing the use of cells by ensuring that they are in a biologic condition able to guarantee their function, and avoiding any potentially dangerous cell transformation. If the product is composed of cells and biomaterials, both components have to be characterized in their identity, purity, safety, and functionality to finally determine their compliance with specific release criteria for a clinical use, as we recently demonstrated for a MSC-based clinical device using collagen scaffolds. ⁴⁰

The first aspect that has to be determined for a GMP-grade validation is the assessment of straightforward conditions to obtain sufficient cell numbers for further manipulations. Therefore, PL has been proposed to replace FBS.^{41,42,43,44,45} In the present study, we have tested 12 random batches of PL, and compared them with FBS. In line with these previous reports, we demonstrated that 5% PL is superior to 10% FBS in terms of cell proliferative capacity. An accelerated cell growth (and therefore a shortening of the culture) can be also further achieved by the use of different culture conditions and cell concentrations at each passage, as we have recently demonstrated. ⁶¹ In a protocol of GMP-grade MSC manipulation, PL medium is to be recommended, since besides avoiding the use of animal-derived materials, an accelerated growth rate diminishes the period of culture. PL lots have to be standardized in their platelet concentration, which is the most relevant element for the release of GFs.^56,57,58 To analyze differences in the release of GFs possibly responsible of accelerated growth rate, ELISA tests were performed to quantify EGF, TGF-β1, PDGF-αα, PDGF-ββ, VEGF, and bFGF, well known in the literature to induce MSC proliferation and, at least for TGF-β1, MSCs further differentiation. We have observed that all PL lots contain a great quantity of the above-mentioned GFs, particularly if compared to human plasma, whose GF levels were undetectable. In line with our findings, previous reports have demonstrated that the use of allogeneic human serum causes MSC growth arrest and ultimately cell death. ⁶² Even though it has been reported that the use of autologous serum may determine an adequate growth of autologous MSCs, ⁶³ the amount of autologous serum necessary for sufficient expansion would exceed the amount a donor could provide. The formulation of PL-release criteria in clinical-grade cell expansions would be desirable, but this objective was not obtained, at least for the GFs tested. The amount of GFs in PL lots is extremely variable and donor dependent, and we could never establish a minimum value for any of the GF tested. The actual contribution of the above-mentioned GF for accelerated growth rate is still a controversial topic, and although we observed a wide difference in GF contents among the 12 PL batches, there was no relevant change in the rate of MSC proliferation from batch to batch, included the ones with lower GF values. These observations are in line with what previously described by Weibrich et al., ³⁰ who analyzed 213 different PRP lots. Certainly, such a great amount of GFs is somehow responsible for accelerated MSC proliferation, but further studies are necessary to determine which ones are truly relevant. Our article is not conclusive at this regard. We actually know that PL will have variable potency for different patients, and this is expected in any cell biology process. For cell expansions we actually use randomly available PL lots from our Transfusion Centre, since GF analysis was not informative in any case and every single lot (included the ones with lower GF values) was excellent for expanding MSCs and preserving their characteristics.

Another relevant aspect that has to be taken into account for clinical application is the possibility that MSCs may undergo malignant transformation, as reported both in humans ⁴⁶ and in mouse models, ⁴⁷ with contradictory results. ⁴⁸ The propensity to transformation has been investigated through several methods. The method most commonly used is the analysis of MSC metaphases. In fact, Rubio et al. ⁴⁶ demonstrated that transformed MSCs always showed chromosome alterations such as trisonomy, tetraploidy, and/or chromosome rearrangements. All our MSC cultures were characterized by a normal karyotype, both in FBS and PL medium. Moreover, MSCs were seeded for growth in methylcellulose and no colony was detected, while transformed MSCs described by Rubio et al. lost contact inhibition and grew in semisolid agar. The same authors also showed that c-myc protein overexpression was linked to transformation in human MSCs. On the contrary, absence or decreased p53 expression is well known to be relevant for the process of neoplastic transformation of various cell types. ⁴⁹ We did not observe any downregulation of p53 or upregulation of c-myc in our expanded cells. The large variability of p53 expression is not due to different lots of PL, because a single lot was used; therefore, we do not predict a correlation of this variability with some of the GFs released by PL.

Besides accelerating growth rate with no trace of tumoral transformation, we also proved that PL medium did not alter the phenotype of expanded MSCs. In accordance with previous reports,^41,43,45 MSCs retained their ability to differentiate into osteogenic and adipogenic lineages, demonstrating that PL does not obstacle the multipotency of these cells. An analysis of gene profile was performed to ensure the quality of the cell product. Since our interest was the osteogenic differentiation of MSCs, several genes implied in this process were analyzed, to exclude the possibility of a dystrophic mineralization. ⁶⁴ The best conditions, the best cell sources, and environment for bone regeneration have been a matter of study for a long time. Differentiation is usually characterized by increased AP activity, RUNX2 expression, and formation of mineralized bone–like nodules containing COL1, OC, ON, and OP. In line with this, Zhang et al. ⁵⁵ recently demonstrated the influence of the ontological and anatomical origin of MSCs, being human fetal MSCs the most promising candidate for bone tissue engineering. Frank et al. demonstrated the heterogeneity across different subpopulations within the same primary culture. ¹² Our results clearly show a very potent effect of PL, combined with a clinical-grade osteogenic induction cocktail, in inducing an accelerated and greater expression of the mentioned genes. Our results in qRT-PCR analysis of osteogenic genes in FBS culture were comparable to literatures,^12,55 while in PL cultures data really overpassed any expected values. Concerning the already reported decrease in osteoblastic potency after culture in the presence of PL, ⁵⁸ these results are not confirmed in our hands, since all the genes were strongly upregulated, as measured by qRT-PCR. We may suppose that different methods of platelet activation (such as thrombin vs. thermic shock) or different platelet final concentration in serum-free medium or different methods of cell culturing could, at least partially, explain the observed differences. Moreover, when using PL, we observed a very high release of TGF-β, which belongs to the TGF-β superfamily, a group of morphogenetic proteins important during the healing of bone. ⁶⁵ Nevertheless, we are well aware that this remains a controversial topic.

When envisaging the use of a GMP-grade combination of MSCs and an osteoconductive support, the final step is ultimately represented by the establishment of a correct 3D system to create a suitable environment to promote cell adhesion and proliferation, and secondly tissue growth and differentiation. HA has been often employed as bone substitute for human clinical application. In vitro–expanded MSCs seeded onto a HA and/or tricalcium phosphate ceramic biomaterials have been used in several experimental animal models,^21–25 but a formal validation of such biological composed devices under GMP conditions has never been provided. The Orthoss^® scaffold is a safe and highly purified deorganified bone mineral, very similar to human cancellous bone. Incorporation and remodeling of the bone substitute material was observed in the absence of any toxic effect.^66,67 With the goal of promoting MSC adhesion to the HA scaffold, we first studied the adequacy of a clinical-grade precoating material (retronectin). Retronectin showed an excellent capacity in permitting cell adhesion. Immunofluorescence studies further demonstrated cell adhesion around the whole perimeter of the scaffold holes. SEM analysis showed that the seeded MSCs could colonize the most internal parts of the scaffold, receiving nutrient from the medium. Even though we do not provide a direct measurement of viable cells, a certain degree of cell viability was indirectly demonstrated by SEM analysis (by the observation of proliferating cells forming confluent cell layers, as shown in Fig. 6), by immunofluorescence (demonstrating integrity of cells by coloration of cytoplasm and nuclei, as shown in Fig. 7), and by the cell count estimation, as shown in Figure 8. Analysis of MSC osteogenic gene expression demonstrated the expression of each single gene tested, even though at lower levels than in monolayers conditions. This could be explained by several factors, including technical difficulties in extracting mRNA from MSCs after mechanical destruction of the scaffold. In addition, since the cells have to reach a suboptimal confluence to differentiate, we can suppose that the variability we found in our scaffolds can somehow affect the success of the assay. It is also likely that our static culture method needs to be further improved to possibly obtain a confluent layer ubiquitously through the whole scaffold. In fact, various investigators^68,69 demonstrated that spinner flask culture methods give beneficial effect on cell growth and differentiation inside the scaffold. Concerning the demonstration of osteogenic potential of the proposed device, we are aware that this is an in vitro characterization and that only an in vivo model could finally prove it. Our data give anyway, in our opinion, a sufficient proof of the tendency of these cells to undergo osteogenic differentiation. The only critical point is represented by OP expression, whose high levels have been described in such a context by our group for the first time. An unexpected single-gene profile does not necessarily mean that the osteogenic process is surely impaired or that this is accompanied by a propensity for a different differentiation process. Only a demonstration in vivo would finally prove if this phenomenon has a significant impact on the process of osteogenesis or it is only due to our peculiar ex vivo culture conditions.

Our results by SEM analysis and confocal microscopy show that Orthoss^® can provide a suitable 3D environment for cell adhesion, permitting adequate distribution, and growth of MSCs and allowing cell differentiation. Nevertheless, we are fully aware that additional issues have to be considered when envisaging the in vivo implant of such a device in sites of bone degeneration. One of this issue is if our proposed 3D device can permit cell migration both inside and outside the product itself. We were not able to measure if any migratory activity in vitro was present in our construct. This issue may be relevant for clinical application, since migration toward surrounding damaged tissues would be desirable. Nevertheless, we can hypothesize that such a phenomenon will most likely occur, possibly induced by the presence of chemokines released by various inflammatory cell types present in the implantation area. The second (probably the most important) issue to be considered is represented by induction of new vessel formation. Different studies have indeed demonstrated that formation of new vessels is necessary for osteogenesis. ²⁷ For such a purpose, we envisage the use of our PL-cultured MSCs on the HA scaffold mashes combined with platelet-gel, at the moment of the surgical treatment of nonunions, since it has been widely described as an optimal source of angiogenic factors. ²⁹

In conclusion, the present report offers a model of an MSC-based bioengineered device, suggesting that Orthoss scaffold loaded with MSCs has excellent characteristics in terms of safety and osteogenic potential, and support its potential use, combined with platelet-gel, in tissue engineering to repair bone defects. We initially envisage the use of such a product in degenerative bone diseases, since they represent so far the main field of application of MSC-based cell therapy in the area of bone local defects. Clinical trials are planned for nonunions and avascular osteonecrosis. Nevertheless, this product could be used for other kinds of nondegenerative bone defects. When such a project will be transferred in a phase I clinical trial, initial information concerning its feasibility, safety profile, and biological functionality will be provided.