Autologous cell-based therapy for treatment of large bone defects: from bench to bedside

At least two major types of endothelial cell lines can be obtained by in vitro culture of mononuclear cells; first, the so-called “endothelial-like cells” or “early EPC” and second, the so-called “outgrowth EPC” or “late EPC”. Early EPC were used in all of our experimental projects. These cells are supposably derived from monocytic/dendritic cells co-expressing some endothelial markers together with leukocyte markers and demonstrating a high VEGF synthesis, some investigators hence designate them as endothelial-like differentiated PBMC [28, 29]. In the following ‘early EPC’ will be referred to as EPC. These cells can be generated in a sufficient amount within 3–5 days from a tenable volume of blood [30]. The contribution of early EPC in forming blood vessels is a matter of debate. Crosby and colleagues have reported that 8.3–11.2% of endothelial cells which developed in sponge-induced granulation tissue over 1 month were derived from circulating hematopoietic progenitor cells [31]. So it has been proposed that early EPC more likely act in a paracrine manner, secreting proangiogenic factors such as VEGF [28]. Own previous work indicated that early EPC are activated after multiple trauma by increased VEGF and TGF-β [32] but are harmed by increased concentrations of TNF-α, IL-1β [33] and activated neutrophils [34].

In contrast, outgrowth EPC or late EPC are characterized by a broad spectrum of endothelial markers including VEGF-R2 and UEA-I-Lectin. They express CD34, lack myeloid markers (CD45) and can be expanded in vitro. It is likely that these cells are generated from bone marrow-derived CD133 + cells [35, 36]. The culture period of late EPC is much longer, compared to that of early EPC. Single colonies of late EPC appear after 3–4 weeks [37], whereas early EPC require only 3–5 days [38, 39]. We observed that late EPC were also activated by musculoskeletal trauma [40], and that migration of late EPC towards injured tissue is impaired in elderly patients probably due to a reduced capability for VEGF synthesis [41].

MSC were primarily described by Friedenstein et al. [42] as plastic-adherent cells or colony forming unit fibroblasts based on their adherence to tissue culture surfaces. MSC own a high proliferative potential and are phenotypically characterized by surface expression of CD71, CD73, CD90 and CD105, and the absence of the leukocyte marker CD45 or markers expressed by hematopoietic stem cells such as CD34 [43]. These cells can be functionally characterized by their potential for trilineage differentiation towards the adipogenic, the chondrogenic or the osteogenetic lineage in dependency from the presence of specific substances in the culture medium. For bone tissue regeneration, MSC combined with an appropriate scaffold have shown to support bone repair [43, 44].

We observed an increased proliferative activity of MSC in patients with multiple trauma and a decreased concentration in the bone marrow of patients who developed an atrophic non-union during our initial research [45].

BMC are a heterogeneous mixture of diverse cell types containing (immature) lymphocytes, (immature) monocytes and progenitor cell populations. A BMC preparation evidentially comprises several subsets of regenerative potential such as (immature) monocytes and hematopoietic stem cells (HSC), a putative source of EPC, and precursors of MSC [46‐49]. Putative MSC precursors can be identified by the expression of the nervous growth factor receptor-1 (CD271) and the absence of the pan leukocyte marker CD45 [50], whereas EPC can develop from CD34/CD133/CD45 expressing cells [51]. MSC precursors are a rare population of cells residing in the bone marrow that were defined by the presence of CD271 expression and, respectively, low or absence of pan leukocyte antigen CD45 expression. Those cells were frequently found in close proximity to CD34 + progenitor cells [50] and possess the potential for trilineage differentiation (adipogenic, chondrogenic, osteogenic potential) [52]. It has been shown furthermore that the CFU-F concentration correlates well with the concentration of those cells within the bone marrow [53] and that approximately 5% of those cells were capable to form CFU-F [50, 53]. It has been demonstrated that BMC support therapeutic effects by improvement of vascularization as exemplarily demonstrated by Jeon et al. [54] using the hind limb ischemia model of the mouse. Transplantation of BMC resulted in significantly increased microvessel density [54]. Interestingly, in cardio-vascular cell transplantation studies, BMCs were mostly applied in large successful studies (Assmus et al.) [46]. BMCs are easily taken by bone marrow aspiration of the iliac crest and processed for further clinical use. The different cells with regenerative potential are shown in Fig. 2.

Stem cells for bone tissue engineering can be harvested from different sources. Human MSC and EPC can be obtained not only from iliac crest bone marrow but can also be isolated from marrow of the femur using a Reamer Irrigator Aspirator (RIA). The application of a Reamer/Irrigator/Aspirator (RIA) system allows the harvest of vital bone marrow from the femur by continuous irrigation and simultaneous aspiration of the irrigation fluid. The irrigation fluid as well as the osseus particles within the irrigation fluid can be harvested using a filter. Actual studies demonstrate that the reaming debris obtained with RIA contains elevated levels of FGF-1, PDGF, IGF-1, TGF-β1, and BMP-1 in comparison with samples obtained from the iliac crest using needle puncture/aspirate technique [55]. Moreover, it was reported recently that human reaming debris is a rich source of multipotent stem cells. The harvested cells exhibit a phenotype and a plasticity commonly attributed to MSC in culture [56]. Own work has shown that in comparison with aspirates obtained from iliac crest RIA aspirates from the femur contained a significantly higher percentage of CD34 + progenitor cells, a significantly higher concentration of MSC and a significantly higher concentration of early EPC. The percentage of late EPC did not differ between both sites. Moreover, the capability of MSC for calcium deposition was significantly enhanced in MSC obtained with RIA [47]. In a subsequently following study, we hypothesized that the harvest procedure influences the osteogenic activity of human MSC rather than the tissue site itself. We generally were able to reproduce that concentration and osteogenic capacity of MSC harvested with RIA is higher compared to MSC from the iliac crest. We observed that the harvest procedure is a critical factor in osteogenesis of MSC in vitro. The altered gene expression and function of femur-derived MSC (RIA) might be due to the harsh isolation procedure [57].

Actually, there is a high demand for cytocompatibility testing, since the effect of the scaffold on cells is not predictable solely based on information about the scaffold’s chemical and physical properties. Therefore, we established a panel of assays that allows us to rate the cytocompatibility of a scaffold for BMC, EPC and MSC in a 96-well plate scale. Our test panel includes the assessment of seeding efficacy, metabolic activity, relative number of adhering cells, evaluation of functional aspects such as the secretion of VEGF and expression of genes relevant for vasculogenesis and osteogenesis.

We observed significant differences of cytocompatibility between different sorts of scaffolds, which were consistently found for each cell type that was analyzed. In particular, the relevance of the physical surface characteristics was demonstrated. It was observed that number, metabolic activity and gene expression of MSC, respectively, EPC, differed significantly when seeded on β-TCP scaffolds being chemically identic but different in their surface topography. Both cell types demonstrated a high adhesion and survival rate on the β-TCP offering a smooth surface, whereas cell number and cell activity rapidly declined on the rough material [58, 59].

Natural materials on the basis of human processed bone material demonstrated a high cytocompatibility for MSC, EPC [58‐60] and BMC [49] that was superior to synthetic materials with regard to initial adherence and long-term survival (Fig. 3).

The importance of certain ions being released from the scaffold for the differentiation, survival and activity of EPC was demonstrated using a composite material developed in our department consisting of a PLA carrier combined with up to 40% bioglass (BG40, CaO-SiO₂–SiO₂ 80 mol-%, CaO 20 mol-%) [61]. BG40 released the most calcium, and improved endothelial differentiation and vitality of EPC best. This effect was mimicked by adding an equivalent amount of calcium to the medium and was diminished in the presence of the calcium chelator, EGTA.

We also analyzed the portability of the in vitro results to the in vivo situation. The general proceeding of those experiments is shown in Fig. 4.

Keeping our initial hypothesis in mind, that the combination of cells with complement properties is more effective for the bone defect healing compared to approaches using single cell sorts, we evaluated the effect of EPC alone or in combination with mesenchymal stem cells (MSC) on the early vascularization and bone healing in our critical size defect model of the athymic rat.

We were able to show that early vascularization after 1 week was significantly improved in the EPC/MSC group and the EPC group. The formation of a primitive vascular plexus was also detectable in the β-TCP, MSC, or autologous bone group, but on a significantly higher level, if EPC were transplanted alone or combined with MSC. The degree of early vascularization correlated well with the release of VEGF into the tissue, suggesting a paracrine effect of the transplanted EPC.

Concomitantly, bone defect healing after 8 weeks was most prominent, if MSC and EPC were transplanted into the bone defect compared to all other groups. Those findings indicated a synergistic effect between EPC and MSC and that the initial stage of neovascularization mediated by EPCs is crucial for complete bone healing in the late phase [62, 63].

The same positive effect of co-transplanted MSC and EPC on bone healing and vascularization was seen in a rat critically sized calvarian defect model using syngenic MSC and EPC seeded on a newly developed scaffold consisting of polylactic acid reinforced with 40% bioglass [61, 64].

Despite their beneficial effects on bone healing, the use of culture expanded cells comes with inherent disadvantages, including regulatory ones. To obtain a sufficient number of cells for clinical use, MSC will require several weeks of expansion in culture, markedly delaying definitive surgical repair of the bone defect. There is some evidence that, during that process, MSC may accumulate genetic alterations, which in turn might increase the risk of cancer [65]. Also, some of the growth factors that are used for EPC differentiation in vitro such as IGF-1 might support transformation of hematopoietic progenitors [66] from which EPC develop [51].

Bone marrow mononuclear cells (BMC) might be a promising alternative to cultured cells, if preliminary data about their osteoinductive properties can be confirmed in humans, specifically also in humans with pathological bone structure. Comparative data regarding the needs of BMC for the adhesion on biomaterials and biocompatibility to various biomaterials are lacking to a large extent. Therefore, we evaluated whether a surface coating would enhance human BMC adhesion and analyze the biocompatibility of three different kinds of biomaterials. β-TCP, demineralized bone matrix (DBM), and bovine cancellous bone (BS) were assessed. The seeding efficacy of BMC on uncoated biomaterials is generally high, although there are differences between these biomaterials. β-TCP and DBM were similar and both superior to BS. Those in vitro results could be generally confirmed using our femur defect model of the rat. Superior bone healing responses of the β-TCP and DBM scaffolds compared to BS were observed suggesting either as suitable materials for spatial restriction of BMC used for regenerative medicine purposes in vivo [49, 67]. Based on those preliminary data, we analyzed the impact of BMC seeded on a β-TCP scaffold in comparison with combined EPC and MSC using the same scaffold in our femur defect model of the male athymic rat in vivo. We observed less chondrocytes and a significantly more advanced bone formation in the BMC and EPC/MSC group in comparison with the control group (β-TCP without cells) after 8 weeks. Concomittantly, biomechanical stability of the defect area was significantly enhanced if BMC and EPC/MSC were implanted compared to control. The degree of new bone formation and biomechanical stability was similar between the BMC and the EPC/MSC group. Furthermore, no tumor formation was found either macroscopically or histologically after 26 weeks of BMC implantation [68].

Based on our promising preclinical results, we established a cell-based bone regeneration procedure applicable in the whole field of bone defects after trauma, tumors, joint arthroplasty and in osteoporotic defects. We hypothesized that transplantation of BMC + β-TCP into a bone defect should be safe, feasible and should promote bone formation and bony bridging of the defect resulting in improved clinical outcomes. The clinical problem of these studies is always that in substantial bone defects mostly the defects are very heterogeneous, and additional problems, such as soft tissue defects, or additional injuries exist. To allow for a rather standardized clinical defect situation, we have chosen the situation of a displaced proximal humerus fracture, thus a metaphyseal defect. Such an approach was proposed by Saxer et al. for their studies on adipose-derived stem cells [69].

Protocols for a German Medicines Law GCP trial were prepared and permissions from the local ethics board [No. 350/12] and the federal authority (PEI) [No. 1769] were obtained for treatment of 10 consecutive, eligible, consenting patients. We generated formal study protocols, including IMPD, and applied for § 40 AMG permission from the PEI for this phase I trial (EudraCT-Nr.:2012-004037-17, Date of registration: 30th of August 2012; Date enrolled first participant: 11th of September 2013). A manufacturing license for tissue procurement acc. to § 20b German Medicines Law and for manufacturing of the advanced therapy medicinal product (ATMP) “BMC2012” acc. to § 13 German Medicines Law, the autologous cell-based study drug, was obtained from the local regulatory agency (Regierungspräsidium Darmstadt). The study was registered in the European Clinical Trial Register as EudraCT No. 2012-004037-17.

After regulatory approval 10 patients were recruited after informed consent and completed follow-up between September 2013 and 2014 and published in 2016 [70].

Criteria for inclusion to this clinical trial were 2-, 3- or 4-fragment fracture (Neer classification), dislocation of ≥ 10 mm between fragments and/or angle of ≥ 45° between fragments and/or dislocation of tuberculum major of ≥ 5 mm, age > 18 years, informed consent for surgery and study participation.

The study was a single-arm uncontrolled study. All patients received cell-based therapy with autologous BMC: open reduction and internal fixation (ORIF) of the fracture, augmentation with composite of an acellular bone graft substitute (β-TCP) and BMC. Concentration of BMC was 1.3 × 10⁶ BMC/ml β-TCP analogous to the prior animal experiments.

Five follow-up visits for clinical and radiological control up to 12 weeks were performed and neither morbidity at the harvest site nor morbidity at the surgical wound site was observed. Furthermore, neither local nor systemic inflammation was noted. All fractures healed within the observation time without secondary dislocation. We conclude that cell therapy with autologous BMC for bone regeneration appeared to be safe and feasible with no drug-related adverse reactions being described to date. The impression of efficacy was given, although the study was not powered nor controlled to detect such [70].

Therefore, a phase IIa-clinical trial was initiated to evaluate the effect of autologous BMC on bone healing. Formal study protocols for a multicentric, open, randomized phase IIa trial (EudraCT-Nr.:2015-001820-51) were generated and approval of the local ethics board [369/15] was obtained. A total of 94 patients distributed prospectively and randomly in a 1:1 relation to verum (BMC) or control group (β-TCP) is estimated and until June 2017, 24 patients have been already enrolled. We expect a study duration of 2–2.5 years and are eager to see if the phase I results can be demonstrated in a prospective randomized trial. The study design is shown in Fig. 5.