A new manufacturing process for allogeneic bone plates based on high hydrostatic pressure – treated granules for jaw augmentation

: Currently used methods for processing allogeneic bone grafts like gamma irradiation are connected with downside of altering the mechanical properties of the graft. As an alternative, high hydrostatic pressure (HHP) leads to an e ﬀ ective devitalization of cells without in ﬂ uencing the bone matrix and its mechanical behavior. To address the clinical application, bone plates were prepared out of HHP-treated bone granules, which are conceivable for augmentations in the jaw region. In order to achieve su ﬃ cient mechanical strength, two di ﬀ erent adhesives were tested. Mechanical characterization by three-point bending tests was performed. Furthermore, analysis regarding cytotoxicity as well as colonization experiments with mesenchymal stem cells were performed to investigate osteoconductive properties of the bone plates. While plates composed of ﬁ brin glue showed better biocompatibility, plates prepared with Loctite ® 408 showed better mechanical properties and could be incorporated in a model application. Regardless of the adhesive, bone plates induced osteogenic di ﬀ erentiation compared to cells cultured without bone plates. Although an adhesive combining both properties would be necessary for later clinical application, the study at hand demonstrates the possibility of producing allogeneic bone plates from HHP-treated granules, which meet the basic requirements for jaw augmentation.


Introduction
The main causes of tooth extractions are caries and periodontitis, besides orthodontically or prosthetically indicated extractions [1]. The loss of teeth is accompanied by various changes in the jaw, for example, loss of bone of the alveolar process. The bone can be reduced by up to 50 % in the buccolingual dimension within 1 year after tooth extraction [2], which can limit subsequent dental implant placement [3]. In order to be able to place implants correctly, bone augmentation is often necessary and can be performed using various materials such as autografts or allografts. Autologous bone is osteoinductive, osteoconductive, and osteogenic and is, therefore, often referred to as the gold standard for augmentations [4]. The Khoury techniquea widely common bone augmentation procedureuses autologous bone blocks harvested from the retromolar region of the mandible, which are attached to the alveolar ridge with screws to augment the jaw. Afterward, the space between the bone block and the alveolar ridge is filled with autologous bone chips [4]. Despite a 95 % success rate, this technique also has drawbacks such as infections or paresthesia at the donor site [3].
Alternatively, commercial allogeneic bone graft like maxgraft ® cortico (botiss biomaterials GmbH, Berlin, Germany) can also be used for augmentation of the jaw. The bone plates have osteoconductive potential and allow augmentations with reduced surgical time and no morbidity at the donor site [5]. Another fundamental advantage of allografts in general is their almost unlimited availability compared to autografts. Yet, the risk for postoperative infections and inflammatory reactions caused by microbiota or proteins remaining in the graft are disadvantageous [6]. This risk can be reduced by graft processing using physical or chemical methods like gamma irradiation, thermal disinfection, or decellularization using sodium dodecyl sulfate (SDS) or strong acids. However, these established procedures often lead to a reduction in the mechanical properties of the allografts [6].
High hydrostatic pressure (HHP) treatment is commonly used in food processing and results in microbial inactivation without affecting taste or vitamins [7,8]. Inactivation of microbes is achieved by inhibition of protein biosynthesis, enzyme deactivation, and via modification of membranes [7]. It has also been shown that drugs such as insulin can be sterilized via HHP without altering molecular integrity [7]. The effect of HHP treatment on mammalian cells has also been investigated in recent years. Using various cell lines, it could be shown that HHP treatment leads to the initiation of apoptosis by activating caspase-3. Among others, this has been investigated in human lymphoblasts, retinal ganglion cell lines, or MEL cells [9][10][11][12]. Different pressures were found to induce apoptosis for different cell lines. For osteoblasts, it was shown that pressures of 250-300 MPa tend to induce apoptosis, while pressures of 450-500 MPa tend to have necrotic effects [13]. Therefore, HHP treatment could be an alternative to achieve effective devitalization of allografts, since, compared to other procedures, the extracellular matrix (ECM) and thus the biomechanical properties remain intact. This has been demonstrated in experiments on arterial vessels, uterine tissue, trabecular, and cortical bone [14][15][16]. The mechanical properties of bone are mainly based on the mineralized components and Diehl et al. could not demonstrate any effect of HHP treatment on collagen type I and proteins of the ECM of bone [17]. Waletzko-Hellwig et al. tested the mechanical behavior of cylinders pressed from bone granules treated with HHP [18]. No effect of the high hydrostatic pressure treatment on the mechanical properties could be measured [18].
The aim of this study was to prepare allogeneic bone plates from HHP-treated granules that can be used in a manner comparable to the Khoury technique or maxgraft ® cortico. As mentioned before, HHP application to structurally intact bone is effective in terms of cell death. Nevertheless, residual cellular components remain in the tissue, which can cause immunological reactions. Due to the structural integrity of the bone, even gentle irrigation steps would be unsuccessful. Thus, HHP treatment of bone granules could have the advantage that downstream rinsing steps to remove cellular debris would be more effective. A pure application as a filling material would be conceivable from this step on. However, this granulate has to be transformed into a specific form with sufficient mechanical properties for insertion on load-bearing regions. To ensure this, the plates must have sufficient mechanical strength. In addition, the plates have to be biocompatible and allow cell colonization and differentiation to enable stable osseointegration postoperatively. For this purpose, granules harvested from human femoral heads were treated at 250 MPa and 20°C for 20 min. These parameters are based on preliminary work by Waletzko-Hellwig et al. and provide effective devitalization without affecting the mechanical properties [18]. Two different adhesives were used to achieve sufficient mechanical strength. The plates were mechanically tested via a three-point bending tests according to previous work [19,20]. Biocompatibility analysis of the adhesives used in the plates were performed by means of MTT assay. To evaluate the differentiation capacity of plates, these were colonized with human adipogenic stem cells from lipoaspirates (hAD-MSC). To simulate clinical use, the plates were also screwed to a mandibular model.

Materials and methods
2.1 Sample preparation, HHP treatment, and production of bone plates Trabecular bone was obtained from femoral heads of patients undergoing a total hip joint replacement. Patient consent and ethical approval from the ethics committee of the University Rostock, Germany were obtained before surgery (ethics approval A 2010-0010). The femoral heads were rinsed using phosphate-buffered saline (PBS) (Sigma Aldrich, Munich, Germany), 1 % penicillin/streptomycin (Sigma Aldrich, Munich, Germany) was added, and samples were stored at −20°C until further use. Prior to processing, the samples were slowly thawed at 4°C. The defrosted bone was sawn into cubes of approximately 5 mm edge length. Bone granules were produced from these cubes using a bone mill with a 2 mm cutting cylinder (USTOMED, Ulrich Storz GmbH & Co. KG, Tuttlingen, Germany). The granules were then subjected to high hydrostatic pressure treatment at 20°C and 250 MPa for 20 min using a high hydrostatic pressure reactor (Dustec Hochdrucktechnik GmbH, Wismar, Germany). For this purpose, the material was placed in PBS. Since the granulate was subsequently very moist, it was freeze-dried at −40°C for 8 h. The granules were filled into a press mold made of aluminum and pressed using a uniaxial testing machine (Z050, Zwick Roell, Software testXpert II, Ulm, Germany) into plates with dimensions of 25 × 10 × 1.7 mm ( Figure 1). To achieve this thickness, 0.9 g of granules and a predefined compression regime was used. Since the plates did not produce a strong bond without further additives, the addition of adhesive was necessary before pressing. To compare different approaches, one group was tested with 200 µL fibrin glue (hereafter FG) (TISSEEL, Baxter International, Deerfield, Illinois, USA) and one with 150 µL Loctite ® 408 (hereafter L408) (Henkel AG & Co. KGaA, Düsseldorf, Germany). The compressive force was applied at 0.5 mm/s. Following the preliminary work by Waletzko-Hellwig et al., a pressing protocol was used in which a setting phase at 0.5 kN (60 s) was applied first [18]. This was followed by an unloading phase at 0.2 kN (10 s) and a loading phase at 10 kN (300 s). Afterward, the plates were removed from the mold and stored refrigerated at 4°C until further use.

In vitro cytotoxicity test
To test the used adhesives for cytotoxicity, in vitro tests with an MTT assay according to DIN EN ISO 10993-5:2009-10 were performed [21]. First, the individual adhesives (L408 and FG) were placed in Dulbecco's Modified Eagle Medium (DMEM, Invitrogen, Darmstadt, Germany) for 24 h (n = 5). Normal human dermal fibroblasts (NHDF) were then treated with 25 µL of each of these samples for 24 h (40,000 cells each). One group received medium without addition of the materials used for the plates, which served as a positive control. After removing the medium, 150 µL of MTT, diluted 1:10 with DMEM, was added to each well and the cells were incubated for 45 min at 37°C/5 % CO 2 . Afterward, supernatants were collected and absorbance was measured at 570 nm wavelength. Viability was calculated as measured by the positive control and determines the cytotoxic potential of the measured sample.

MSC colonization and evaluation of differentiation
In order to evaluate differentiation capacity of human adipogene mesenchymal stem cells (hAD-MSCs) cultivated with manufactured plates, cells were isolated and expanded according to a standardized protocol [22]. Per each group, 12 plates were seeded with 50,000 hAD-MSCs. Cells, seeded on the bottom of the well plate without bone samples, served as a control. They were incubated at 37°C/ 5 % CO 2 and nourished in stem cell culture medium containing F12 and IMDM (Gibco™, Thermo Fisher Scientific Inc., Waltham, MA, USA) for 7, 14, and 21 days. The supernatants and the corresponding samples were collected every 7 days and medium was changed. Four plates per group and per time period were analyzed regarding gene expression (ALP, OPG, BGLAP, Table 1) and protein expression (osteoprotegerin -OPG, transforming growth factor β -TGF-β, bone morphogenic protein 2 -BMP-2) by flow cytometry. For gene expression analysis, plates were collected and digested overnight with 1 mL collagenase A (Roche Diagnostics GmbH, Mannheim, Germany). Afterward, samples were centrifuged at 1000×g for 5 min. Supernatants were removed and pellets were resuspended with 400 µL Lysis Buffer (innuPREP RNA Mini Kit, Analytik Jena, Jena, Germany) and incubated for 10 min at room temperature. Control cells, incubated without bone samples, were treated with the same amount of lysis buffer for the same  time period directly in the well plates without prior digestions.
Afterward, samples were either stored at −80°C for further use or RNA was isolated directly according to manufacturer's protocol (innuPREP RNA Mini Kit, Analytik Jena, Jena, Germany). The amount of RNA was quantified with the Tecan Infinite Reader Pro using a Nano Quant™ plate (both: Tecan Trading AG, Maennodorf, Switzerland). For transcription of RNA into cDNA, 100 µL RNA per sample was applied at RNAse free water was added to a total volume of 10 µL. To each sample, 10 µL MasterMix was added and reverse transcription was performed with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) related to manufacturer's protocol (10 min at 25°C, 120 min at 37°C, 15 s at 85°C). Afterward, 20 µL RNAse free water was added to each sample and cDNA was stored at −20°C for further use. For quantitative PCR, to 1 µL sample a master mix containing 3 µL RNAse free water, 0.5 µL forward primer, 0.5 µL reverse primer, and 5 µL innuMIX qPCR DS Green was added. Samples were pipetted in duplicates and qPCR was performed using a qTOWER running the following protocol: 2 min at 95°C, 40 cycles for 5 s at 95°C, and 25 s at 60°C. Results were analyzed using the ΔΔCt method, whereas a cycle threshold was set to a Ct-value of 30. First, Ct-values were related to the housekeeping gene gapDH (ΔCt = Ct target −Ct gapDH ). Afterward, ΔΔCt was calculated relating the ΔCt to the Ct of unstimulated control (ΔΔCt = ΔCt -ΔCt control ).
Supernatants used for protein quantification were stored at −20°C. At the day of analysis, they were gently thawed and protein content was quantified using a customdesigned LEGENDplex™ (BioLegend ® , San Diego, CA, USA). Samples were used undiluted with a total amount of 18 µL per sample. Further preparation was performed according to manufacturer's protocol and measurement was performed using the BD FACS Verse™ (BD Bioscience, Heidelberg, Germany) supplied by the Core Facility of Cell Sorting and Cell Analysis (University Medical Center Rostock, Germany). Measured values were related to the total protein content, which was analyzed using the Qubit Protein Assay Kit ant the Quibit 1.0 (both: Invitrogen, Waltham, MA, USA) according to manufacturer's protocol.

Mechanical characterization
To evaluate mechanical properties of the plates, three-point bending tests were performed at room temperature using a uniaxial testing machine (Z1.0, Zwick Roell, Software testXpert II, Ulm, Germany). The radius of the fins was 1.5 mm and the span width 20 mm. A preload of 0.1 N was applied at a cross-head speed of 0.5 mm/s. After preloading, the plates (L408 n = 15, FG and control n = 14) were loaded until fracture with 0.05 m/s, while force and displacement were recorded. These parameters are based on previous studies and represent a physiological loading [23,24]. The measured force and displacement were used to determine the flexural strength and the flexural rigidity. The experimental setup can be seen in Figure 2. Native bone plates of the same size, which were sawn from trabecular bone, served as a control.

Application to a jaw model
To simulate the possibility of clinical application, two holes were drilled in each of the plates using a 1.5 mm dental drill for 2 mm screws (KLS Martin Group, Tuttlingen, Germany). Through these holes, the plates were fixed in the jaw model ( Figure 3) with 2 mm screws (Modus ® Trauma 2.0, Medartis AG, Basel, Switzerland).

Evaluation of the results and statistics
For the mechanical characterization, the flexural strength was calculated using the ratio of the maximum moment at fracture and the axial section modulus of the plates. The flexural rigidity was determined from the slope of the linear region of the force-displacement curve. This determination was performed using Microsoft Excel 2016 (Microsoft Corporation, Redmond, Washington, USA).
Statistical analyses were performed using GraphPad Prism version 7 (GraphPad Software, San Diego, CA, USA). Before testing for statistical significances, all variables were evaluated for normal distribution via Shapiro-Wilk test. Depending on the presence or the absence of normal distribution, additional analyses were conducted using oneway ANOVA, two-way ANOVA, and Bonferroni's as a posthoc test, respectively. p-Values <0.05 were considered as significant. The results are presented as box whiskers plots or bar graphs.

Preparation of bone plates
Using the described protocol, uniform bone plates could be reproducibly manufactured. While the width and length were determined by the mold, a uniform thickness of 1.7 mm ± 0.1 mm could be achieved. Figure 3 shows an example of both, plates held by adhesives, and native control plates sawn from trabecular bone. The plates of the control group were subjectively very rigid, while the FG plates were rather soft and flexible. The L408 plates appeared less rigid in manual handling than the plates of the control group but appeared harder than those of the FG group.

In vitro cytotoxicity test
To test the biocompatibility of the adhesives used, a cytotoxicity test with MTT was performed. The results are depicted in Figure 4. The fibrin glue achieved 100 % cell activity. Cells incubated with the L408 showed significantly less activity on average (p = 0.0024) but with 82.5 % more than the cytotoxicity limit of 70 %.
NHDF were treated with supernatants of L408 and FG and incubated at 37°C and 5 % CO 2 for 24 h. Results are shown as box plots with median and interquartile ranges from 25 to 75 %. Statistical analyses were performed using an unpaired t-test with Bonferroni's multiple comparison test (n = 5). Statistical significances between groups and time points: **p ≤ 0.01.

hAD-MSC colonization and evaluation of differentiation
Gene expression analysis were performed to evaluate the capacity of plates to induce osteogenic differentiation of hAD-MSCs. Therefore, the amount of mRNA transcripts of characteristic genes occurring during osteogenic differentiation was analyzed ( Figure 5). ALP expression decreased over time from 7 to 21 days of incubation. This is reflected in differences between time points and was observed regardless of the adhesive used for production of plates (7 vs. 21 days: FG p = 0.0004; L408 p = 0.074). While significant overexpression of ALP was detected at day 7 compared to the control samples, no differences were observed at day 14 and 21. Furthermore, hAD-MSCs incubated with FG plates showed higher expression of ALP than cells seeded on L408 plates. The expression of BGLAP also showed a reduction  over time. However, in this case, no differences were observed neither between the adhesives nor comparing the expression to the control group. Strikingly, BGLAP mRNA transcripts were not expressed significantly above the control level already at day 7 of incubation. Furthermore, expression dropped below the control. Thus, BGLAP differed from the other analyzed genes ALP and OPG, which did not fall below the control level. OPG, as well as ALP, already showed significant overexpression compared to the control on day 7. This was observed for both, FG samples and L408 samples. Furthermore, as for ALP, hAD-MSCs incubated with FG plates showed a higher expression of OPG than hAD-MSCs seeded on L408 plates. Over time, expression of OPG decreased in both groups on the control level but increased again on day 21 of incubation. Plates were pressed according to the described protocol using fibrin glue or Loctite ® 408. Samples were colonized with hAD-MSCs and incubated at 37°C and 5 % CO 2 for 7, 14, and 21 days. Gene expression data are depicted as 2 (−ΔΔCt) values and hAD-MSCs expanded on the bottom on a cell culture plate served as a control (y = 1). Statistical analysis was performed by using two-way ANOVA with Bonferroni's multiple comparison test (n = 4). Statistical significances between groups and time points: *p ≤ 0.05; ***p ≤ 0.001; statistical significance compared to control #p ≤ 0.05; ##p ≤ 0.01; ####p ≤ 0.0001; tendencies 0.1 ≤ p ≤ 0.05 are depicted as §.
In order to determine secretion of proteins characteristic for bone formation, cell culture medium was analyzed regarding representative mediators ( Figure 6). In the FG group, OPG levels were detected at every time point. Thereby, the level initially decreased after 14 days compared to 7 days of but increased at day 21. This trend did not show statistical significance at high standard deviation. OPG was present in the supernatants of the L408 group only in the earlier two time periods, and here each showed comparable values to those of FG. In contrast, OPG was not detected in the control at any time.
In the early period, TGF-β could be detected in both experimental groups and in the control. The value for the L408 group with 0.76 ng/mg total protein was significantly higher than in the FG group (0.06 ng/mg total protein) or the control (0.03 ng/mg total protein). At the intermediate time point, the TGF-β level within the L408 group decreased significantly to 0.13 ng/mg total protein, whereas the level in the FG group increased slightly to 0.16 ng/mg total protein. At the late measured time point, TGF-β could only be detected in the L408 group, where the value further decreased to 0.03 ng/mg total protein compared with the earlier time points.
BMP-2 was detected in all samples at every time point. The peak was either in the early (L408 group) or in the intermediate period (FG group, control). At the late time point, the measured concentration of BMP-2 decreased in all groups. Compared with low BMP-2 concentrations in the control (<0.1 ng mg −1 total protein), significantly higher levels were measured in the experimental groups (FG up to 0.45 ng mg −1 total protein, L408 up to 0.31 ng mg −1 total protein).
Plates were pressed according to the described protocol using fibrin glue or Loctite ® 408. Samples were colonized with hAD-MSCs and incubated at 37°C and 5 % CO 2 for 7, 14, and 21 days. hAD-MSCs expanded on the bottom on a cell culture plate served as a control. Crossed-out boxesnot detected. Statistical analyses were performed by using twoway ANOVA with Bonferroni's multiple comparison test (n = 4). Statistical significances between groups and time points: *p ≤ 0.05; **p ≤ 0.01; ****p ≤ 0.0001.

Mechanical characterization
The results of mechanical characterization are summarized in Figure 7. For flexural rigidity, native controls showed significantly higher values than L408 (p = 0.0282) and FG groups (p < 0.0001). Compared to FG, L408 showed significantly higher flexural rigidity (p = 0.0444); however, this difference was less clear than the variance between FG and native bone. These observations can also be transferred to the investigation of the flexural strength (Figure 7b). Again, significant differences between native control and FG (p < 0.0001) or L408 (p = 0.0189) could be detected. Direct comparison of L408 and FG showed significant higher values for L408 (p = 0.0083), but to a lesser extent than in the comparison between native control and FG.
Mechanical properties were tested using a three-point bending test. Data are shown as box plots with median and

Application on a jaw model
It was possible to drill two holes in the L408 plates without any problems. The subsequent attachment to the model using screws could also be carried out in the sense of the application (Figure 8). The plates could be firmly attached to the jaw without allowing any movement. The FG plates, on the other hand, could not be attached to the jaw model. Due to insufficient mechanical strength, the drill holes tore out, so that stable fixation could not be achieved with the help of the screws.

Discussion
After tooth loss, jaw augmentation is often necessary to create a sufficient large bone site for implantation [3]. Various techniques exist to widen a jaw. Among others, both the Khoury technique, which works with autologous bone blocks, and the allogenic bone plates maxgraft ® cortico are established procedures [3][4][5]. But a major disadvantage of autogenous bone is its limited availability and donor side morbidities, while allogenic materials have their mechanical properties altered by established preparation procedures [6,25].
Therefore, HHP, a method originally derived from the food industry, could be an alternative method to process allogeneic bone grafts without changing the mechanical properties [7]. By applying high hydrostatic pressure, effective devitalization of cells in the tissue can be achieved without affecting the components of the ECM [13][14][15][16][17].
This study tested a new way to produce allogeneic bone plates that meet the basic requirements for clinical use in the field of jaw augmentation. For this purpose, granules of trabecular bone were treated with HHP and subsequently pressed into bone plates. Two different adhesives were used to provide sufficient mechanical strength. To verify the possibility of clinical use, biocompatibility studies and a mechanical characterization were performed.
The cytotoxicity test using MTT showed no cytotoxic potential for fibrin glue (FG). This was to be expected, as the adhesive has been regularly used in clinical application for years and the manufacturer reports excellent tolerability and no evidence of cytotoxicity to fibroblasts in vitro [26]. Besides fibrin glue, also L408 was chosen as a comparative glue, although the manufacturer does not make any statement regarding the biocompatibility of this product [27]. Alternatively, the biocompatible Loctite 4011 (L4011) could have been chosen, but according to the manufacturers data sheet, this glue has less strong adhesive properties than L408. This was confirmed in initial tests, as plates made with L4011 fell apart, eliminating this adhesive from further studies. For biocompatibility testing of L408, in this study cell activity between 61 % and 100 % could be measured at an average of 82.5 %. According to DIN EN ISO 10993-5:2009-10, a substance has cytotoxicity potential if the value falls below the limit of 70 % cell activity [21]. Although the mean value was 82.5 %, some of the individual samples are below this limit. In addition to the cytotoxic effect, this observation could also result from different proliferation rates or various activity levels of the cells. A clear statement as to whether L408 is biocompatible or cytotoxic is, therefore, difficult to make on the basis of the available data. Compared to the FG, however, the L408 is significantly less biocompatible.
In addition to biocompatibility, allografts should promote osteogenic differentiation to allow the migration and differentiation of stem cells into osteoprogenitor cells and subsequently form new bone. This was investigated by analyzing the expression of genes and the secretion of proteins characteristic for differentiation processes. Predominantly, overexpression of genes and proteins that promote bone formation compared to the control were detected in both groups. ALP, BMP-2, and TGF-β are considered markers of the early stages of bone formation as they promote the development of stem cells through osteoblast precursor cells to active osteoblasts [28][29][30][31]. The three mentioned mediators were detected at early time points in this study on hand, suggesting an induction of osteogenic differentiation. Furthermore, it was shown, that the expression of ALP mRNA transcripts dropped to the level of the control group over time in both groups after high levels on day 7. This observation as well as the detection of high levels of BMP-2 and TGF-β at 7 days and 14 days of incubation followed by a down-regulation also indicate the induction of osteogenesis. Regarding the expression of BMP-2 and TGF-β, differences between FG and L408 varied here and did not show a clear tendency toward one of the two adhesives.
OPG and BGLAP are known to be expressed by active osteoblasts and, in contrast to the previously mentioned mediators, are considered rather late markers of bone formation [32,33]. For these, there were no differences between the two adhesives, however, compared to the control. Whereas BGLAP was initially measured at similar concentrations to the control, it dropped below the control level after 21 days. BGLAP is a representative of noncollagenous bone matrix proteins, and its expression is apparently not stimulated by the incubation with bone plates. A possible reason for this could be that there is no need for BGLAP expression as the matrix is stable due to adhesives and the associated strength of the specimens. It is also possible that hAD-MSCs were not well enough differentiated to allow the detection of BGLAP. This is also supported by a study in which only the incubation of osteoblasts with additives led to the detection of BGLAP [34].
However, due to the detection of OPG, on both, protein and gene expression level, the differentiation of hAD-MSCs into osteoblasts can be assumed. OPG is responsible for the preservation of bone matrix by inhibiting osteoclast maturation [35]. Interestingly, OPG protein increased in the FG group after 21 days, but in the L408 group, it was only detectable after 7 and 14 days. On the other hand, an overexpression for both adhesives compared to the control was detected at gene expression level on day 21. Due to the time offset of gene expression and protein secretion, the incubation period might have had to be extended in order to detect OPG also in the L408 group at protein level. While the genes and proteins that typically occur during early differentiation to osteoblasts were also increasingly measured in the early time periods in this study, the results for the late markers were not as consistent. In principle, it should also be noted here that bone granules themselves may still contain proteins characteristic for bone and that those are secreted over time. The amount of the secretion can also be influenced by the adhesives. It is conceivable that L408, which holds the plates together more strongly, also holds back proteins.
Even, if not all markers appeared completely as expected, these results show that plates with both L408 and FG significantly promote osteogenic differentiation of hAD-MSCs compared with the control. This effect can be attributed to the osteoinductive properties of the allogeneic bone, which in turn result from the growth factors present in the bone [36,37]. In order to better assess the in situ situation, further studies should also focus on histological staining. This would allow statements to be made on cell distribution and also deposition such as osteocalcin or calcium, which would contribute to a comprehensive overall picture. In addition, repeating the experiments with bone marrow stem cells (BM-MSCs) would be useful, since some studies showed a reduced differentiation capacity of AD-MSCs compared to BM-MSCs [38]. Cultivation of plates with BM-MSCs could thus lead to clearer results.
In addition to biocompatibility, the bone plates must provide sufficient mechanical stability to allow intraoperative processing and implantation. Both the three-point bending tests and the attempt to attach the plates to a jaw model showed that the L408 plates are more stable than the FG plates. In addition, the comparison with native bone plates in the bending tests showed that these are significantly more stable than both fabricated plates. This can be explained by the fact that the trabecular arrangement of the plates is not destroyed, which is the case for the L408 and FG plates due to the grinding of the bone. Since the plates do not have to carry loads for jaw augmentations [5], lower stability does not mean lower suitability. In fact, the lower flexural rigidity could even be an advantage, as the plates are thus more flexible, easier to model, and better able to adapt to the individual shapes of the jaw. It is particularly important for the application that the plates can be firmly screwed to the model and that they allow the gap to be filled with bone augmentation materials so that postoperative remodeling with the formation of autogenous bone can take place. However, the failed attempt to attach the FG plates to a jaw model showed that these plates are not suitable for the intended application. The L408 plates also showed lower strength than the native control in the bending tests but attached well to the jaw model. Although the transferability of these in vitro experiments to actual intended use is limited, the L408 plates appear to meet the mechanical requirements for clinical use and the subsequent implantation of common dental implants [39]. Since native bone plates are more stable than plates made of granules, it seems reasonable to use native plates. However, since HHP only has a devitalizing effect and not a decellularizing effect, this is not a viable alternative. Immunogenic cell residues would remain in the allograft and could thus lead to inflammation. By using bone granules instead, it is possible to rinse the tissue after HHP treatment and thus remove cellular debris.
Limitations of the experimental study include the simplified design of the test setups. Although the bending tests provide an indication of the mechanical properties of the plates, they cannot fully represent actual clinical use, where multiaxial forces act on the grafts. Similarly, the cytotoxicity tests and the determination of gene and protein expression were performed under simplified conditions. Healing of grafts in the human body involves additional cell types as osteoclasts and is a complex process that can only be represented in a highly simplified way by the experiments performed [40].

Conclusions
In summary, the FG plates are biocompatible and the L408 plates seem to have good mechanical properties for clinical use. Therefore, to produce plates from HHP-treated bone granules, an adhesive would be needed that behaves mechanically like the L408 but is as biocompatible as the FG. Further research will be needed to solidify these statements. The transferability of these in vitro experiments to a clinical application is limited. Firstly, an adhesive should be found that combines the properties of the FG and L408. For further studies, cell experiments would be needed that not only investigate bone-formation processes but also osteoclasts, which are responsible for bone resorption processes [41]. If such experiments are successful, bone plates could be tested in an animal study, such as a calvarial bone defect model [42]. Although this would provide insight into the mechanical forces acting during jaw augmentations, such a model could reveal important results regarding intraoperative handling as well as biocompatibility.