Cell-Based Therapy by Autologous Bone Marrow-Derived Mononuclear Cells for Bone Augmentation of Plate-Stabilized Proximal Humeral Fractures: A Multicentric, Randomized, Open Phase IIa study

Abstract

Proximal humerus fractures are common in an aging population. The standard operative treatment is open reduction internal fixation (ORIF) using an angular stable plate. However, this procedure has complications such as a relatively high rate of secondary dislocation, humeral head necrosis or nonunion caused by delayed bony consolidation. Autologous bone marrow mononuclear cells (BMC) combined with a β-TCP scaffold could support bone healing and is considered clinically safe.

This multicentric, randomized, open phase IIa clinical trial (Clinical Trials. Gov Identifier: NCT02803177, Eudra CT No: 2015-001820-51) evaluated whether autologous BMC with β-TCP in addition to ORIF reduces the incidence of secondary dislocations in patients with proximal humerus fracture. Ninty-four patients equally divided between verum group (BMC+β-TCP) and control group (ß-TCP only) were targeted and calculated. At the time of planned interim evaluation, ie, enrolment of 56 patients, no statistical difference in secondary dislocations or complications was demonstrated in either group after an observation period of 12 weeks. Radiographic bone healing and DASH score to determine shoulder function were comparable between both groups. Bone marrow harvest and BMC transplantation did not result in any severe adverse events. Therefore, the study was terminated after the interim analysis, as no other result could be expected. From the study results, it can be concluded that the application of autologous BMC is well tolerated, and bone healing can be achieved. Augmentation of bone defects with β-TCP could be shown to be feasible and might be considered in other clinical situations.

Graphical Abstract

Graphical Abstract.

Lessons Learned.

• The application of autologous BMC is well tolerated, and bone healing can be achieved.

• Augmentation of bone defects with β-TCP could be shown to be feasible and might be considered in other clinical situations.

Significance Statement.

In this first prospective clinical study an autologous cell transplantation on a scaffold in a clinically relevant fracture situation, the proximal humerus fracture, we aimed to demonstrate a possible translational pathway from in vitro over experimental in vivo data to the clinical situation. The observed bony consolidation of the fractures did not differ between the treatment groups which might be attributed to the use of bone replacement material in both groups. It can be clearly concluded that BMC2012 did not have any negative effect on bone healing. The absence of SAE demonstrated the good tolerability of autologous cell therapy with BMC2012.

Introduction

Large bone defects caused by severe trauma, nonunion fractures, or tumor resection usually fail to heal spontaneously due to impaired regeneration capacity.¹

Autologous bone grafts can be used for bone augmentation, their limitations and complications include donor site morbidity, questionable biological quality, insufficient quantity, and extended surgery time. Allogeneic bone grafts must be heat inactivated before use, limiting their use to space holding, stabilization, and provision of osteoconductive matrix.^2-4

In recent years, bone tissue engineering has increasingly begun to combine osteogenic and endothelial cells, synthetic osteoconductive scaffolds (eg, β-tricalciumphosphate, β-TCP), and biological factors, which has offered a promising alternative approach to treat large bone defects and minimize or eliminate the above-listed limitations or costly complications.^5,6

Previously, we demonstrated that cell-based therapy using extensively cultivated stem cells (mesenchymal stromal cells, MSC, endothelial progenitor cells, EPC) implanted on a β-TCP scaffold into a femoral large-sized segmental bone defect in rats leads to improved vascularization and new bone formation in the defect site.^5-7 But these methods are afflicted with disadvantages like long duration of cell culture, possible risk for genetic alterations or contamination with pathogens and hence delay of definitive surgery, biological safety issues, high costs, and not yet established clinically.⁸

In contrast, BMC (bone marrow-derived mononuclear cells) can be harvested and reintroduced to the patient within hours which is more compatible with the clinical requirement for rapid fracture repositioning. BMC have been used in cardiology without severe side effects in clinical use due to their apparent regenerative potential and safety profile.⁹ In our own preclinical work, BMC on synthetic matrix of a certain chemical quality and pore size provided highly beneficial effects on bone healing response probably by improving vascularization in an athymic rat model of critical size defects. The histological quality and the mechanical strength of femur defects treated with BMC were qualitatively comparable to those defects receiving cultured MSC and EPC^10,11 and markedly better than bone defects augmented with synthetic matrix without cells.

Based on those results, we aimed to establish a cells plus matrix-based bone regeneration procedure applicable in the whole field of bone defects after trauma, tumors, arthroplasty and in osteoporotic defects.

After permission were obtained from the local Ethics board and the federal autority (PEI), Marzi et al performed a phase I clinical trial of autologous bone marrow mononuclear cell therapy for proximal humerus fractures.¹² Based on the literature, the risk of secondary dislocation (varus collapse of the humeral head) for osteopenic proximal humerus fracture was estimated at 15%-48% due to delayed bony consolidation (complications up to 30%) and is one of the most often fractures with the indicated problems.¹³

A prior clinical phase 1 study had confirmed safety and feasibility of an augmentation using the autologous cell product (BMC2012) in combination with an angle stable fixation (Philos plate) for the therapy of proximal humeral fractures. The cell product BMC2012 consists of preoperatively isolated BMC seeded in situ onto β-TCP in a density of 1.3 x 10⁶ BMC/ml β-TCP. Although a complication rate of 20%-30% in bone healing had been expected on the basis of published data described, none of the 10 consecutively included patients developed a secondary dislocation within the study period of 3 months.¹⁴

Knowing the high safety of BMC in the phase I trial¹² and the estimated beneficial effects on bone healing, a prospective, clinical phase IIa trial was planned, applied for, and conducted to test the efficacy of a cell-based augmentation with preoperatively isolated autologous BMC seeded onto β-TCP in combination with an angle stable fixation for the therapy of proximal humeral fractures. The primary outcome parameter was the frequency of secondary loss of the reduction result after 12 weeks, which is defined as varus dislocation of ≥20° of the head-shaft angle in the true antero-posterior (a.p.) radiograph compared with the primary reduction result, as well as secondary perforation of screws. Furthermore, a shoulder function score was assessed, bone healing was evaluated in a blinded fashion, and safety parameters were determined.

Patients and Methods

Objectives

This was a clinical trial phase IIa to test in a controlled trial the efficacy of bone augmentation with preoperatively isolated autologous BMC cells seeded onto β-TCP in combination with angle stable fixation for the therapy in large bone defects of proximal humerus fractures.

Ethics and Regulatory Affairs

A manufacturing license for tissue procurement according to §20b German Medicines Law and for manufacturing of the advanced therapy medicinal product (ATMP) “BMC2012” (see below) according to §13 German Medicines Law, the autologous cell-based study drug, was obtained from the local regulatory agency (Regierungspräsidium Darmstadt). Protocols for a German Medicines Law GCP trial were prepared and permissions from the local Ethics board (No. 369/15) and the federal autority (PEI) (No. 2554/01) were obtained for treatment of 94 consecutive, eligible, consenting patients. The study was registered in the European Clinical Trial Register as EudraCT No. 2015-001820-51. Informed consent was signed by all patients enrolled. No study-specific X-rays were performed.

Experimental Group/Control Group

The 47 patients of the verum group were supposed to receive the cell-based therapy with implantation of BMC2012 during the surgical treatment of proximal humerus fractures.

The 47 patients of the control group should receive only the cell-free bone graft substitute β-TCP. The case number calculation was based on published data on the expected complication rate and the nonoccurrence of complications in the phase I safety study of BMC augmentation for proximal humerus fracture.¹² Further details are described in the statistics section. Patient acquisition took place at 2 centers (Center 1: University Hospital Frankfurt; Center 2: Berufsgenossenschaftliche Unfallklinik (BG) Frankfurt). The 24 hour-online randomization was performed by an independent institution (Fraunhofer Institute for Translational Medicine and Pharmacology, ITMP, Frankfurt, Germany) after enrolment of the patients. Each study site was randomized separately with the ratio of 1:1.

Inclusion/Exclusion Criteria

The inclusion and exclusion criteria were summarized in Table 1.

Table 1.

Summary of inclusion and exclusion criteria.

Inclusion criteria	Patients between the ages of ≥50 and ≤90 with a proximal humerus fracture
	Indication for an open reduction and internal stabilization with a proximal fixed-angle plate of the humerus (PHILOS, Synthes, Oberdorf, Switzerland)
	2-, 3- or 4-fragment Neer fracture
	Dislocation of ≥10 mm between the fragments and/or angle of ≥45° between the fragments and/or dislocation of the tubercle majus from ≥5 mm
	Negative pregnancy test in premenopausal women
	Ability to understand the nature, scope and significance of the clinical trial
	Signed consent form for surgery and study participation
Exclusion criteria	Pregnancy and lactation
	Luxation fracture
	Known mental illness, which makes cooperation considerably more difficult (eg, dementia, schizophrenia, severe depression)
	Incapacitated patients
	Pathological fractures caused by other underlying diseases
	Fracture-related nerve damage
	Tumor disease with adjuvant therapy or treatment within the last 3 months (eg, chemotherapy, radiotherapy), untreated tumour diseases
	Participation in a clinical trial within the last 3 months before inclusion in this trial

Preparation and Application of the Autologous Cell Product BMC2012

On the day prior to surgery, 50 mL bone marrow were aseptically aspirated from the posterior iliac crest under local anesthesia in regulator-approved intervention rooms. In addition, 27 mL of peripheral blood were aseptically collected into endotoxin-free “no additive” vials, for manufacturing of patient-derived serum. Bone marrow aspirate and blood were transported immediately under standardized conditions (20°C) to the Department of Transfusion Medicine and DRK Blutspendedienst Baden-Wurttemberg-Hessen, Frankfurt. The blood samples in the serum vacuettes remained with the bone marrow aspirate at all times and were used to prepare autologous serum under GMP conditions. For this purpose, clotted blood was centrifuged at 2800g for 15 minutes. The serum was aseptically aspirated (clean room class A in B) and used for final drug formulation.

BMC preparation was performed under full GMP conditions in the certified facility of the Department of Transfusion Medicine of the Goethe University as described in reference.¹² Briefly, BM aspirate was diluted in saline and subjected to Ficoll (Lonza, Verviers, Belgium) density gradient centrifugation. The interphase cells were carefully collected, pooled, washed, and resuspended in X-vivo (Lonza) containing 20% v/v autologous serum. All open and semi-open steps were performed in a class A in B, all other steps in a class B environment. BMC were counted, diluted to a final concentration of 1.3 × 10⁶/mL suspension media and aseptically transferred to a cryostorage bag (Miltenyi Biotec, Bergisch-Gladbach, Germany). The final product consisted of 12 mL BMC suspension, and it was stored at room temperature until use. Total leukocytes (WBC) were enumerated using the Sysmex XT1800 hemacytometer (Norderstedt, Germany). Content of putative progenitor cells (CD271+/CD73+/CD45− [putative MSC]; CD45+/CD34+/CD133+/VEGFR2+ [putative EPC]; CD45+/CD34+/CD133+ [putative hematopoietic stem progenitor cells, HSPC]) was determined by single-platform flow cytometry using Trucount counting beads (Becton-Dickinson, Heidelberg, Germany) and a standard setup 3-laser FACS Canto II (Becton Dickinson). A sample for sterility assessment (available postrelease) was taken by overfilling the final product bag by 1 mL and withdrawing 1 mL, half of which was subsequently inoculated each onto aerobic and anaerobic BacT/Alert bottles (BioMerieux, Nürtingen, Germany). Quality check further included assessment of vitality (7AAD) and serological testing on a companion tube of patient blood for Hepatitis A, B, C, HIV, and Treponema pallidum. All release-critical assays used were validated according to guidelines laid out in the European Pharmacopoiea. The time between quality assessment and implantation was approximately 24 hours but could be in seldom cases 48 hours. The shelf life of the product BMC2012 was assessed in a prior study.¹² The vitality was assessed over 72 hours, and only this time point vitality was significantly decreased compared to 24 hours, whereas the vitality at 24 hours and 48 hours was equal to the vitality assessed directly after production.

The final investigational product BMC2012 consisted of 16 × 10⁶ ± 20% nucleated blood cells (white blood cells, WBCs) in 12 mL X-Vivo 10 medium with 20% autologous serum. The bone defect was filled with a known volume of β-TCP-granules (1.4-2.8 mm, Chronos, DePuy-Synthes, Dubendorf, Switzerland), and immediately an equal volume of BMC2012 was carefully combined with the β-TCP in situ by gentle pipetting. The resulting cell density was 1.33 × 10⁶ BMC per mL β-TCP-scaffold. The volume of the injected BMC-suspension equals the volume of implanted β-TCP. The mean defect size for this fracture type ranges usually from 4-5 cm³. Therefore, the volume of BMC-suspension ranged between 4-5 mL.

Patients assigned to the control group received only the β-TCP granules for bone defect filling during surgery.

Course of Study and Clinical Data Collection

The study design anticipated a 3-month observation period postoperatively and included a total of 6 visits (V0-V5). All clinical trial data were transferred to an electronic case report form (eCRF, IMTP, Frankfurt, Germany). The first visit (V0) identified study participants and included initial screening, study inclusion, and randomization. The patients were not blinded to their treatment. Patients of the verum group received a bone marrow punction. Patients of the control group did not received a sham intervention because of ethical reasons. The participants of the study were informed about the results of the randomization and their group assignment. During the second visit (V1), which occurred a maximum of 48 hours before surgery, patients in the verum group underwent bone marrow aspiration by iliac crest puncture to provide the starting material for generation of the autologous BMC. Furthermore, peripheral blood was drawn for the generation of autologous serum as a component of the BMC2012 product, for infectious serology and safety laboratory (control group only safety laboratory). The third visit (V2) took place on the day of surgery. The cell product BMC2012 was implanted in the verum group, additional medication and adverse events (AE) were assessed, and the safety laboratory was collected and a radiograph was taken, which was not part of the study but was used as a clinical control. The fourth visit (V3) took place within the first postoperative week, the safety laboratory was taken, and additional medication and AEs were recorded. Another radiograph was obtained for clinical control, but not as a study component. At the fifth visit (V4), which occurred 6 weeks (± 1 week) postoperatively, examinations were performed analogous to visit V3. At the sixth visit (V5), which completed the study and was conducted at 12 weeks (±2 weeks) postoperatively, examinations were again performed analogous to visits V3 and V4. In addition, shoulder function was determined using the DASH score. The study design is also illustrated in Fig. 1.

Figure 1.

Workflow of the BMC2012 clinical phase IIa trial.

Primary Endpoint

The primary outcome parameter was osseous healing of fracture after 12 weeks assessed by radiological evaluation. The frequency of secondary loss of the reduction result after 12 weeks, which was defined as varus dislocation of ≥20° of the head-shaft angle α, ∆α in the true a.p. radiograph compared to the primary reduction result was analyzed (Fig. 2).

Figure 2.

Determination of head shaft angle α in a true a.p. radiograph. The angular increase between the time of final surgical treatment and at the three-month postoperative time point can be used as a measure of the success of the treatment. An angle difference of α>20° within three months is rated as a complication.

Analysis of Fracture Healing

The analysis of fracture healing as well as the analysis of the secondary perforation of screws was done radiologically by evaluating the bony superstructure in the true a.p and the outlet-view X-ray. The X-rays were also taken to determine the fracture position and implant location and to detect screw breakage, osteonecrosis, and implant failure (loosening or fracture). To detect secondary dislocation, the head-shaft angle was measured in the true a.p. radiograph. To detect secondary screw perforation/cut out, the distance between the screw tip and the joint surface was measured for the screw closest to the surface.

The true a.p. and outlet-view X-rays of the shoulder were performed at visit 2-5 according to clinical standards in order to evaluate fracture site and implant position as well as detect screw-cutting-out, osteonecrosis, pseudarthrosis, and loosening of implants and definitive bone healing. Due to fracture-immanent possible secondary varus dislocation, we measured head-shaft-angle in true a.p. X-ray at visits 3, 4, and 5. For this, a line from the upper to the lower limit of joint surface was drawn (A-B line), then a perpendicular line to A-B line through the center of the humeral head (C-D line). The angle alpha between this line and the bisectual line of humeral shaft (E-F line) was measured as head-shaft-angle. Secondary dislocation was defined as a secondary loss of reposition result of ≥20° of head-shaft-angle in the true a.p. view (Fig. 2).

Radiographs from representative patients (Con, Verum) for visits V3 and V5 were presented in the revised Fig. 3. The angle α is included in each X-ray image. We decided to omit X-rays of visit 4, as they do not provide any further information.

Figure 3.

Representative radiographs of surgically treated proximal humerus fractures fixed with angle-stable plates postoperatively (visit V3) and after 12 weeks (visit V5) in one patient of the control group (β-TCP) and one patient of the verum group (β-TCP+BMC). No major changes (Δα > 20°) in the angle α between V3 and V5 occurred in either group.

Functional Outcome

Shoulder function was tested by DASH (Disabilities of the Arm, Shoulder and Hand) -Score at V5 (12 weeks postoperatively).¹⁵ DASH is a “self report” questionnaire in which the patient answered 30 questions about his current condition, which record the global function of the upper extremity within the last week. The DASH-Score allows a statement regarding function, symptoms, and special activity (athletes, musicians). A DASH-score of zero is a result with an optimal function without limitation. A DASH score of 100 means a maximal limitation.

Systemic Inflammation, Local Side Effects

Systemic inflammation WBC, CRP, IL-6, PCT were measured by standard clinical diagnostic laboratory. Furthermore, fever (>38.5°C, longer than 2 days) and local infection or side effects (inflammation, wound healing disorder) were recorded.

The concomitant medication (visits 0-5) and AEs (visits 2-5) were registered. Specifically for the compatibility of BMC2012, also extraction morbidity of the bone marrow was queried.

Statistics

A case number calculation was performed for the primary outcome variable to determine the necessary number of patients. A power of 80% and a significance level of 5% were assumed. With an assumed frequency of 30% complications (Δα > 20°) in the control group as based on the literature¹³ and a frequency of 0% in the verum group and an assumed drop-out rate of 10% due to patient age and comorbidities, the group size for the dichotomous study variable was n = 47 per group (total patient number equals n = 94).

The Chi-square test was used to compare the frequency of secondary dislocations (Δα > 20°) in the treatment groups (β-TCP only, BMC2012+β-TCP) as well as for all other dichotomic variables.

Physical measures (eg, mediators in serum) were analysed quantitatively using the nonparametric Wilcoxon Mann Whitney U-test. A 2-sided P-value <.05 was significant. The data tables for the statistical analyses were generated from the eCRF by the IMTP. Data in tables were presented as median values, 25% (Q1) and 75% (Q3) quartiles.

Statistical analyses were performed after extensive consultation and under supervision by biostatisticians from the Fraunhofer Institute for Translational Medicine and Pharmacology, (ITMP, Frankfurt, Germany) using Bias 11.0 software (Epsilon-Verlag, Darmstadt, Germany).

Results

General Study Course and Study Termination After Intermediate Analysis

We generated formal study protocols, including IMPD, and applied for §40 AMG permission from the PEI for this Clinical trial phase-IIa (EudraCT-Nr.: 2015-001820-51). Ninety-four patients were planned, 58 patients were included, 1 drop-out, and 1 screening failure occurred. The study period was from July 2, 2016 (date of first enrolment) until June 6, 2019 (date of last completed) and at that time point 56 patients could be analyzed.

After regulatory confirmation and approval by the ethics committee an interim analysis after the scheduled completion time of the study showed that there is no significant difference between the 2 treatments with regard to the primary outcome parameter (frequency of secondary dislocation). The study was terminated after review and evaluation of the data by an independent statistician. For ethical reasons, it was recommended to terminate the study, as it would not have been possible to achieve a significant difference even if the remaining patients would have been recruited.

The study was, therefore, defined as completed within the scheduled time, but with fewer patients (56 instead of 94 patients) than planned. It has to be stated that no negative medical aspects were associated with the treatment of the autologous BMC preparation in all of the treated patients.

Demographic Data

A 2-armed, randomized, placebo-controlled phase IIa clinical trial was conducted. Over a period of 42 months, 58 patients were enrolled in the study at the 2 participating study centers, and of these 56 were considered for evaluation. There was one dropout in the control group, and one patient was excluded due to screening failure. The demographic data of the study participants are summarized in Table 2.

Table 2.

Basic demographic data of study participants.

	Control (n = 28)	Verum (n = 28)	P-value
Age	64.5 (Q1: 60.3/Q3: 71.8)	68.0 (Q1: 60.5/Q3: 74.3)	.477
Male/female	8/20	9/19	1.00
Dropout	1	0
Screening failure	1

Quality Assessment of the Cell Product BMC2012

No bacterial contamination of the individual cell preparations was observed. The serological testing of patient blood for Hepatitis A, B, C, HIV, and Treponema pallidum did not reveal any signs of infection.

The quality assessment of BMC2012 included a multicolor flow cytometric analysis to determine major cell populations with putative effect on bone healing processes. The median vitality was 97.4%. While the concentrations of HSPC and immature HSPC were comparatively high, the fraction of putative MSC precursors and especially the population of putative EPC was very low. The measured values are listed in Table 3.

Table 3.

Viability and cellular composition of BMC. Values are presented as median, Q1 (25% quartile) and Q3 (75% quartile). All values are presented as absolute numbers [n/µL], except the viability, which is presented as percentage of BMC.

Parameter	Identified as	Concentration [n/µL], (median)	Variance (Q1/Q3)
Viability	7AAD negative	97.4%	96.6%/98.1%
HSPC	CD34+/CD45+	16.3	13.6/20.2
Immature HSPC	CD34+/CD133+/CD45+	6.2	2.4/9.1
Putative MSC	CD271+/CD73+/CD45−	0.11	0.06/0.22
Putative EPC	CD45+/CD34+/CD133+/VEGFR2+	0.006	0/0.01
Monocytes	Via FSC/SSC characteristics	170	112.5/225

No Significant Differences in Bone Healing

The aim was to investigate whether transplantation of BMC2012 into the bone defect leads to a reduction in the complication rate of impair bone healing in patients with proximal humerus fracture. Median values were presented in the following. The first-mentioned value refers to the control group and the second-mentioned value to the verum group. The primary outcome parameter was the osseous healing of fracture after 12 weeks. The frequency of secondary loss of the reduction result after 12 weeks, defined as an increase in the angle difference Δα > 20° between postoperative status and after 12 weeks of healing was analyzed. The assumed frequency of this complication, based on literature references and own preliminary work, was up to 30%.^13,14 The central hypothesis was that the frequency of these complications would be significantly lower in the verum group. After evaluation of the data, this hypothesis cannot be answered unequivocally, as no difference in the primary study endpoint (angle difference Δα > 20°) was found (Fig. 3).

Therefore, the originally planned statistical analysis of the frequency distribution was not performed. The direct statistical comparison of the measured angles α between the control and verum group resulted in significantly higher values directly postoperatively (visit 3) in the control group (139° vs. 131°, P < .05), but not at visit 5, 12 weeks after surgery (133° vs. 125°, P = .09). A statistically significant reduction in the angle α between visit 3 (postoperative) and visit 5 (12 weeks postoperative) was observed in both the control group (β-TCP only) and the verum group (β-TCP+BMC2012) (control, V3: 139° vs. V5: 133°, P < .05; verum, V3: 131° vs. V5: 125°, P < .05). The extent of the decrease (angle difference Δα) did not differ between control and verum group, this was 6° for control and 6° for the verum group, respectively (P = .326, Fig. 4).

Figure 4.

Angle α in the control (Con) and verum (Ver) group immediately after surgical treatment of the fracture and implantation of the cell preparation (V3) and after 12 weeks healing time (V5). *P < .05.

There were no significant differences in screw positioning and surface of the humeral head between the postoperative X-ray (visit 3) to the end of the study (visit 5).

Functional analysis using the DASH Score also revealed no significant differences between the control group (Supplementary Material S1).

Safety Aspects

Bone marrow aspiration was performed without complications in all patients in the verum group. No complications at the donor site were registered during the further course of the study. No severe adverse events (SAE) were observed. Analysis of serum samples did not show significant differences in inflammatory parameters (IL-6, PCT, CRP, leukocytes) between both groups preoperatively, postoperatively and to the follow-up rounds (Supplementary Material S2).

Discussion

In the present multicentric, randomized open phase IIa clinical trial, the efficacy of augmentation with an autologous cell transplant, BMC, on the complication rate of a plate-stabilized proximal humerus fracture was for the first time prospectively and randomized investigated. This investigator initiated clinical ATMP study was a demanding study for an orthopedic trauma department but could be managed with a high dedication of all participants and without any problems according to the state authorities and monitoring agencies. However, despite the observed high tolerability of the cell therapy, the hypothesis that the additional administration of BMC2012 significantly reduces the complication rate could not be demonstrated. The radiological findings of the healing results, as well as functional outcome (DASH score) showed no significant differences between verum and control group. Despite the very supportive experimental data of the beneficial effect of bone healing by autologous BMC,¹¹ this effect could not be demonstrated in this study.

To this day, in the clinical setting, the large majority of patients with bone defects still and often receive complete cancellous bone grafts from the iliac crest, which has the disadvantage of donor site morbidity and only provides limited amounts of material.¹⁶ Other approaches such as the use of nonviable scaffolds cannot demonstrate a sufficient biological activity and guidance of bone healing.¹⁷ Alternatively, vital bone derived products such as autologous bone marrow aspirate concentrate (BMAC) can be obtained intraoperatively, mixed with a bone replacement material and transplanted into the bone defect as reviewed in reference.¹⁸ The authors concluded that the processing of BMAC needs to be optimized to obtain a qualitatively standardizable final product that has predictable efficacy.¹⁸ However, recent comparative animal studies indicate that transplantation of purified cells may be more effective for the treatment of bone defects at various sites compared with BMAC. For example, Ardjomandi et al reported that ficollated bone marrow, which is basically equivalent to BMC, has a significantly higher MSC concentration compared with BMAC. In vivo, significantly higher bone growth is achieved in an ovine sinus defect model by transplantation of BMC and bone substitute compared to BMAC. The authors attribute this to a higher concentration of CFU-F in the BMC preparation compared to BMAC. Another problem with BMAC is that there are no randomized studies and mostly small series and case reports.¹⁸ However, it is not clear from this publication at which concentration the BMC were inserted into the defect.¹⁹

The use of different regenerative cell types, or cell mixtures, to improve bone healing is the subject of several clinical studies as well. Various cell types have been and are being investigated. First and foremost are MSC, which have had a comparatively variable assessment of suitability in the past. The reservations against the use of MSC as a cell therapy agent were based on the assumption that in the course of several weeks of culture expansion, genomic mutations accumulate, which in turn might increase the risk of malignant diseases.²⁰ However, recent evidence indicates that the use of MSC can be considered safe as reviewed by Elgaz and colleagues.²¹

The safety but also the effectiveness of autologous MSC for the treatment of large bone defects could be impressively demonstrated in a past clinical study. Gomez-Barrena and colleagues treated pseudarthroses of the femur, tibia, and humerus with autologous, in vitro expanded MSC from bone marrow aspirate. Clinically, bony consolidation was observed in 92% of cases (26 of 28 cases) after a 12-month observation period. In contrast to the study presented here, the MSCs were used at a 100-200-fold higher concentration. Furthermore, the combination of the cell therapeutic agent was not performed in situ, as in the present study, but ex vivo via a special syringe system, which allowed homogeneous loading of the bone substitute material with the autologous MSCs.^22,23

The safety of BMC is implied by the phase I clinical trial¹² and again by the study presented here. A total of 38 patients had proximal humeral fractures treated with BMC and no severe adverse events were observed during the 12-week observation period. Furthermore, no significant morbidity was recorded, either systemically or locally, as a result of bone marrow harvesting.

Systemically, the inflammatory markers CRP, IL-6, leukocyte count, and procalcitonin remained at the comparable level of the control group. The increase in values at visit 3 (first week postoperatively) in both groups compared to the preoperative, and the later visits V4 and V5 is attributed to the surgical trauma²⁴ and comparable for both groups.

The high safety profile of BMC has also been demonstrated in other studies. Both in clinical trials in cardiology and in our comparatively small phase I trial on the safety of BMC in the treatment of proximal humerus fractures, no side effects were observed that were presumed related to the applied cell product. In the meantime, there are also long-term data available for BMC therapy of arterial occlusive disease over 10 years after implantation. The BMC showed a high clinical efficacy. Over the 10-year period, the proportion of amputation-free patients was significantly higher in the BMC group (70%) compared to the control groups. The authors also reported that no severe adverse events related to cell transplantation occurred.^12,25-27

BMCs have a major advantage over MSCs in that they are readily available, and thus can also be used, for example, to treat an acute bone defect. Their production is much simpler and less expensive than that of MSCs due to the lack of an expansion phase. Autologous BMC are considered safe.

Besides their proven safety BMC exerted comparable efficacy to MSCs in the critical bone defect model of the rat femur.¹¹ Based on the flow cytometric results, the number of implanted MSC progenitors is in the range of approximately 2 × 10⁴-3 × 10⁴ cells per cm³ of implanted β-TCP scaffold. This concentration is very low compared to other clinical approaches in which up to 1 × 10⁷ MSC are implanted,^22,23 and it is reasonable to assume that the MSC in BMC are unlikely to contribute significantly to bone healing which might be mediated by other cell populations. Cell types with proven regenerative properties represented in significant concentrations in BMC include HSPC (CD34+) and monocytes. As we have previously conclusively demonstrated, the therapeutic effect of BMC largely resides within the CD14+ population, whereas depletion of CD34+ cells was inconsequential for the bone-inducing activity of BMC. Concentrations of both cell species are several orders of magnitude higher than MSC precursors, so relevant numbers are likely to be implanted into the defect. HSPC subpopulations exhibit EPC properties and thus can potentially support vascularization.^28,29 An animal study also demonstrated that these cells in high local concentrations can support bone defect healing.^30,31 However, our own studies using a 5 mm femoral defect in the athymic nude rat also showed that depletion of CD34+ cells from human BMC had little effect on the bone anabolic efficacy of BMC. A significantly greater contribution to bone healing is made by CD14 positive monocytes. This was demonstrated by a related depletion experiment. When CD14 positive monocytes were depleted from the human BMC preparation prior to implantation, the bone healing outcome in the athymic rat femoral defect model deteriorated significantly in terms of biomechanical loading capacity and new bone tissue formation.³² One might hypothesize that the concentration of monocytes in the aforementioned experimental study was significantly higher than in the clinical study presented here, which may explain the lack of effectiveness of BMC2012 preparations. The differences in monocyte concentrations could possibly be due to the age of the bone marrow donors. In the animal studies,^11,32 BMC from anonymized bone marrow donors were used, which generally have a significantly younger age with a median of 40 years³³ than the patients in the study presented here with a median age of 68 years.

That cell mixtures in combination with bone substitute materials support the formation of new bone in proximal humerus fractures was demonstrated by Saxer et al in a clinical safety study with 8 patients.³⁴ However, there were relevant differences compared to the study presented here, apart from the same fracture type and patient age. Thus, Saxer et al used cells of the stromal vascular fraction (SVF) of adipose tissue. The cell concentration used was higher by a factor of 30, and flow cytometric analysis of the cell mixture showed a proportion of 21% hematopoietic progenitor cells, which is also significantly higher compared with the work presented here. Furthermore, a ceramic material with a fibrin hydrogel was used as the support. Another aspect is that the cell mixture was prepared directly in the operating room and was, therefore, considerably fresher than the cells in the study presented here, where there was a period of at least 24 hours to a maximum of 48 hours between collection and transplantation. The effect of SVF cells on new bone formation was examined histologically in 6 of 8 patients on biopsies taken 8 weeks after surgery as part of a change of fixator. In 5 of the 6 cases, new bone tissue formation was recorded in the pores of the bone graft substitute. Because there was no control group in this safety study, this effect cannot be formally attributed unequivocally to the implanted SVF cells. However, data also presented from an animal study using the athymic rat suggest this.³⁴ Furthermore, it cannot be excluded that in the study presented here, new bone formation also occurred at the level of the implanted BMC-loaded granules. However, since no adequate analyses were performed in this regard, this is in the area of speculation.

Another important aspect is the absence of complications in the control group in the present study. In contrast to the studies reporting a complication rate of up to 30%,^13,14 in the study presented here the defect was filled with β-TCP granules in addition to plate osteosynthesis. Although this does not correspond to the general standard of treatment, it has been already the subject of many studies. Biermann et al summarized the available studies in a review article. For instance, there are biomechanical studies using synthetic or cadaveric bone and retrospective patient studies in which the additional insertion of a bone substitute into the fracture area in proximal humerus fracture was investigated.³⁵ Despite a heterogeneity of the materials used, such as calcium phosphate cements, calcium triphosphate paste or hydroxyapatite, the biomechanics are usually improved after implantation of the material and also the retrospective clinical studies show that the complication rate is significantly reduced by application of ORIF and additional augmentation with autologous or allogeneic bone material but also with synthetic materials such as calcium sulfate or calcium phosphate.³⁵ Thus, there are arguments that in the study presented here, the additional augmentation of the bone defect with a bone graft substitute alone has a positive effect on the healing outcome, which may not benefit significantly more from the addition of BMC.

Conclusion

In this first prospective clinical study an autologous cell transplantation on a scaffold in a clinically relevant fracture situation, the proximal humerus fracture, we aimed to demonstrate a possible translational pathway from in vitro over experimental in vivo data to the clinical situation, which is possible in an academic setting over a time period of 10 years.³⁶ The observed bony consolidation of the fractures did not differ between the treatment groups, which might be attributed to the use of bone replacement material in both groups. It could not be established whether cell therapy in the defect areas of a proximal humerus fracture with BMC2012 can reduce the complication rate of fracture healing in this anatomic region. This might be due to a possible low rate of transplanted cells³¹ or the selection of the clinical defect situation. It might be that diaphyseal bone defects might benefit more than the selected metaphysal defects. However, based on the data available, it can be clearly concluded that BMC2012 did not have any negative effect on bone healing. The absence of SAEs, as well as the inconspicuous course of the inflammation parameters, demonstrated the good tolerability of autologous cell therapy with BMC2012. Thus, autologous cell transplantations, or cell products, still may be an option to follow in order to improve bone healing in the future.

Funding

The authors would like to thank the LOEWE Center for Cell and Gene Therapy Frankfurt funded by “Hessian Ministry of Higher Education, Research and the Arts” for financial assistance (funding reference number: III L 5 - 518/17.004).

Conflict of Interest

The authors declared no potential conflict of interests.

Author Contributions

C.S.: Conception and design, financial support, manuscript writing, Final approval of manuscript. C.N., R.H., R.V.: Collection and/or assembly of data, Provision of study material or patients. D.H.: Conception and design, Collection and/or assembly of data, Data analysis and interpretation, Manuscript writing, Final approval of manuscript. M.B., N.F., V.M.: Administrative support, Collection and/or assembly of data. K.H.S.H., J.F.: Collection and/or assembly of data. H.B.: Provision of study material or patients, Collection and/or assembly of data, Manuscript writing, Final approval of manuscript. E.S.: Administrative support. I.M.: Conception and design, financial support, Collection and/or assembly of data, Data analysis and interpretation, Final approval of manuscript.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.