The Effect of Structural Design on Mechanical Properties and Cellular Response of Additive Manufactured Titanium Scaffolds

Restoration of segmental defects in long bones remains a challenging task in orthopedic surgery. Although autologous bone is still the ‘Gold Standard’ because of its high biocompatibility, it has nevertheless been associated with several disadvantages. Consequently, artificial materials, such as calcium phosphate and titanium, have been considered for the treatment of bone defects. In the present study, the mechanical properties of three different scaffold designs were investigated. The scaffolds were made of titanium alloy (Ti6Al4V), fabricated by means of an additive manufacturing process with defined pore geometry and porosities of approximately 70%. Two scaffolds exhibited rectangular struts, orientated in the direction of loading. The struts for the third scaffold were orientated diagonal to the load direction, and featured a circular cross-section. Material properties were calculated from stress-strain relationships under axial compression testing. In vitro cell testing was undertaken with human osteoblasts on scaffolds fabricated using the same manufacturing process. Although the scaffolds exhibited different strut geometry, the mechanical properties of ultimate compressive strength were similar (145–164 MPa) and in the range of human cortical bone. Test results for elastic modulus revealed values between 3.7 and 6.7 GPa. In vitro testing demonstrated proliferation and spreading of bone cells on the scaffold surface.


Introduction
Large segmental defects in long bones still represent a challenging task in orthopedic surgery.These defects can be caused by fracture, tumor or infection, with varying severity [1][2][3].For large defects (critical size defects), regeneration cannot be accomplished by the patient's body alone.Further assistance is needed, which is provided by filling the defect with bone scaffold materials.
Autologous material is still deemed the "Gold Standard" for treating this kind of defect, because of its high biocompatibility [4].However, grafting of autologous material is associated with problems, which include donor site morbidity, limited availability and the necessity of a second surgery with further consequences for the patient [5,6].Therefore, artificial materials, such as calcium phosphate and metals, have been investigated with regard to the treatment of bone defects, and are used with increasingly frequency [7,8].Metallic materials like titanium and its alloys have performed particularly well in clinical applications, are commonly available, and can be manufactured in a wide range of scaffold designs.For sufficient bone ingrowth into the scaffold, cell biological, in addition to biomechanical, properties play a crucial role in the initial stability of the bone-implant composite system.For bone ingrowth, an open-porous structure and adequate pore sizes must be guaranteed [9].
Furthermore, mismatch of the mechanical properties, between scaffold and surrounding tissue, can lead to stress shielding around the scaffold and subsequently inhibit tissue ingrowth or cause implant loosening [10,11].Therefore, the mechanical properties of the implants acting as scaffolds for bone ingrowth should be adapted to the mechanical properties of the surrounding tissue [12].
In addition to different materials, open-porous structures (i.e., porosity) can reduce the mismatch between the scaffold and surrounding tissue.Porosity, pore size and interconnecting pores are also essential for sufficient bone ingrowth [13,14].
Porosity of the scaffolds can either be achieved by foam-like structures with irregular pore geometry and stochastic pore distribution with varying pore sizes [9,15,16], or by lattice structures with regular geometry and controlled pore sizes [17][18][19].The latter is primarily constructed using additive manufacturing (AM) processes, which also allow the fabrication of complete implants with smooth or structured surfaces [20][21][22].
AM processes facilitate a diverse variety of scaffold designs, in contrast to classical fabrication techniques, like casting or forging.Scaffolds are fabricated layer by layer from powder particles, melted or sintered at defined areas with a high-energy beam (laser or electron beam) [17,23].This offers the ability to control of the architecture, and thus the mechanical properties of the scaffolds can be directly driven by the designs, which are virtually limitless.
Nevertheless, the mechanical properties of open-porous scaffolds are often correlated with their porosity, with controversial findings about the validity [24][25][26].
Furthermore, in addition to the mechanical properties, cell biological compatibility also plays an important role in bone ingrowth into the implant material.Open-porous implants made of titanium or its alloys have already demonstrated their ability for osteointegration in vitro [23,27,28] and have also been successfully used in in vivo studies [18].Nevertheless, the geometry and size of the pores influence the spreading and proliferation of bone cells [29,30].
In order to assess the suitability of open-porous titanium scaffolds with controlled pore geometry as bone scaffolds, mechanical tests were performed on three different scaffold designs characterized by similar porosities.Stress-strain relation ultimate compressive strength, and ultimate compressive strain) For in vitro testing, human osteoblasts were seeded on (SLM)-fabricated scaffolds, which had order to analyze the migration capacity of cells within the scaffo pro-collagen synthesis ability of the

Generating the Open-Porous Scaffolds
The scaffold designs were generated Corporation, Concord, Massachusetts, USA).Three different designs featuring varying structural shapes and strut orientat 14.8 and 4.0 mm, respectively.The first two designs exhibited struts with a rectangular cross orientated vertically.The strut width and height were 400 and 800 between the two layers was 1.3 mm scaffold, the structural shapes in the vertical struts were shifted by half the strut third scaffold were orientated diagonally to the vertical axis and exhibited a circular cross a diameter of approximately 300 µm.was approximately 550 × 550 µm.

Fabrication of the Scaffolds
Based on the CAD data, the scaffolds (n selective laser melting process (SLM solutions GmbH, Lübeck, Germany) from titanium powder (Ti6Al4V).The manufactured scaffolds are shown in strain relationships, as well as mechanical properties (structural modulus, and ultimate compressive strain), were analyzed.human osteoblasts were seeded onto the surfaces of , which had pore geometries similar to the mechanically tested scaffolds order to analyze the migration capacity of cells within the scaffold pores.Furthermore, the type of the bone cells was determined.
Porous Scaffolds generated using CAD software (SolidWorks 2008 Corporation, Concord, Massachusetts, USA).Three different designs were structural shapes and strut orientations.The height and diameter respectively.The first two designs exhibited struts with a rectangular cross trut width and height were 400 and 800 µm, respectively.mm, and the pore size was 800 × 800 µm. in the x-z and y-z planes were identical.For the second scaffold, the vertical struts were shifted by half the strut height (i.e., 400 µm) in the x-z plane.The struts for the third scaffold were orientated diagonally to the vertical axis and exhibited a circular cross µm.The distance between the layers was 1.2 models of the three investigated structures: (a) scaffold with rectangular struts aligned in the vertical direction; (b) scaffold with shifted strut alignment in diagonally-orientated circular struts.
Based on the CAD data, the scaffolds (n = 3 for each design) were fabricated by means of a selective laser melting process (SLM solutions GmbH, Lübeck, Germany) from titanium powder (Ti6Al4V).The manufactured scaffolds are shown in Figure 2. 1338 ies (structural modulus, analyzed. selective laser melting similar to the mechanically tested scaffolds, in ld pores.Furthermore, the type I CAD software (SolidWorks 2008; SolidWorks were created (Figure 1), eight and diameter of the samples were respectively.The first two designs exhibited struts with a rectangular cross-section, respectively.The distance µm.In the case of the first identical.For the second scaffold, the plane.The struts for the third scaffold were orientated diagonally to the vertical axis and exhibited a circular cross-section with istance between the layers was 1.2 mm, and the pore size caffold with rectangular caffold with shifted strut alignment in x-z plane; fabricated by means of a selective laser melting process (SLM solutions GmbH, Lübeck, Germany) from titanium powder

Calculating Scaffold Porosity
Porosity values for the CAD scaffolds with idealized geometry manufactured scaffolds (AMS), were calculated according to the following equations ‫ݕݐ݅ݏݎܲ‬ where V str is the volume of the CAD scaffold struts and outer periphery.‫ݕݐ݅ݏݎܲ‬ where ρ 0 is the density of non-porous Ti6Al4V (4.43 scaffolds, calculated using the weight and volume of the scaffolds.

Axial Compression Testing
All scaffolds were mechanically tested compression tests until mechanical failure were Zwick Roell, Ulm, Germany) with a traverse velocity of 1.0 applied load and displacement were The elastic modulus for each scaffold the testing machine, together with according to Equation (3): where F R is the applied load, l 0 is the initial length, shortening of the scaffold during testing.compressive strain were calculated from Scaffold with identical strut design in x-z and y-z plane; (c) scaffold with diagonal struts.
Porosity values for the CAD scaffolds with idealized geometry, and also were calculated according to the following equations is the volume of the CAD scaffold struts and V cyl is the overall volume enclosed by the porous Ti6Al4V (4.43 g/cm³) and ρ sc is the density of the manufactured weight and volume of the scaffolds.
mechanically tested, in order to determine their mechanical properties.Axial until mechanical failure were carried out using a universal testing machine ( Zwick Roell, Ulm, Germany) with a traverse velocity of 1.0 mm/min for all sca applied load and displacement were continuously recorded during testing.
scaffold was calculated, using the applied load and displacement of together with the geometric parameters of the manufactured test samples is the initial length, A is the initial cross-sectional area shortening of the scaffold during testing.In addition, ultimate compressive strength and ultimate compressive strain were calculated from the stress-strain relationship.
sectional area, and ∆l is the , ultimate compressive strength and ultimate

Cell Seeding on SLM Scaffolds
In order to determine the biological suitability of the SLM-fabricated scaffolds, the migration of human osteoblasts was analyzed.Isolation and cultivation followed the procedure described by Jonitz et al. [31].
Scaffolds for in vitro testing were made as discs, using the same SLM manufacturing process as the scaffolds used for mechanical testing.These in vitro scaffolds were 5 mm in height and 30 mm in diameter, with a rectangular pore size of approximately 700 × 700 µm in all three spatial directions.After cleaning in an ultrasonic bath, scaffolds were sterilized in an autoclave.In order to determine cell proliferation on a scaffold, two discs were put together.As a result, the complete scaffold module was composed of two discs with four different planes (plane 1: superior; planes 2 and 3: intermediate; plane 4: inferior).Human osteoblasts (4 × 10 5 cells) were then seeded point-wise as a 10 µL cell suspension onto the top surface of this two-piece scaffold, which now had a total height of 10 mm.

In Vitro Characterization
Characterization of the mitochondrial activity of the bone cells and the synthesis of pro-collagen type I was determined using a WST-1 assay and an enzyme-linked immunosorbent assay, respectively.Furthermore, cell viability was analyzed by means of the LIVE/DEAD assay.All procedures are described in detail by Jonitz et al. [31].

Scaffold Porosity
The calculated porosities for the CAD scaffolds and manufactured samples are listed in Table 1.All scaffolds exhibited a porosity of approximately 70%.Deviation between the idealized structure and the corresponding manufactured structure was less than 2%.Moreover, variations within each of the three manufactured structures were less than 1%.

Mechanical Behavior
Apparent stress-strain relationships for the scaffolds were calculated on the basis of the nominal cross-sectional areas of the scaffolds (Figure 3).Consequently, apparent stress did not reflect the true stress within the scaffold (i.e., the struts).For each type of scaffold, only one graph has been plotted as an example.Both types of scaffold with rectangular response with a small region of yielding, followed by a sudden decrease due to the failure of the vertical struts within from the scaffold, shifted slightly sidewa the remaining scaffolds did not present The third type of scaffold did not exhibit any plastic yielding.Instead, failur occurred stepwise.After each failure the stress the structural framework along the vertical axis.Furthermore, small parts of the material broke away.Subsequently, damage occurred wi

Mechanical Properties
The mechanical properties of the scaffolds ultimate compressive strain) were derived from calculated from the slope of the elastic response maximum stress prior to failure.ultimate compressive strength.The re Engineering stress-strain relationships for the three types of manufactured ne graph is plotted for each type of scaffold, as an example: ) rectangular struts with shifted strut alignment; and (3) diagonally of the mechanical failure for each type of scaffold Both types of scaffold with rectangular-shaped struts (Figure 3, lines 1 and 2) response with a small region of yielding, followed by a sudden decrease at approximately 6% strain due to the failure of the vertical struts within a single layer.These layers were completely separated slightly sideways and could be easily removed after unloading.Nevertheless, present any indication of plastic deformation.The third type of scaffold did not exhibit any plastic yielding.Instead, failur occurred stepwise.After each failure the stress increased again.The scaffold the structural framework along the vertical axis.Furthermore, small parts of the material broke away.Subsequently, damage occurred within the entire scaffold during testing.he mechanical properties of the scaffolds (i.e., elastic modulus, ultimate compressive strength ) were derived from the stress-strain relationship the slope of the elastic response.Ultimate compressive strength was defined as the failure.Ultimate compressive strain was the strain corresponding to the ultimate compressive strength.The resultant values are listed in Table 2.

1341
for the three types of manufactured , as an example: (1) rectangular diagonally-orientated of the mechanical failure for each type of scaffold are shown to 1 and 2) exhibited an elastic at approximately 6% strain single layer.These layers were completely separated easily removed after unloading.Nevertheless, The third type of scaffold did not exhibit any plastic yielding.Instead, failure of the scaffold caffold displayed disruption of the structural framework along the vertical axis.Furthermore, small parts of the material broke away.elastic modulus, ultimate compressive strength, and strain relationships.Elastic modulus was ltimate compressive strength was defined as the ltimate compressive strain was the strain corresponding to the

In Vitro Properties
To analyze the viability of human osteoblasts within the scaffold, cells were seeded on superior plane 1.One day after seeding, a lot of cells were observed on plane living cells were detected on plane detectable on planes 1-3.These cells also formed densely populated surface on both planes.It was also striking that a lot of dead cells could be determined within unpopulated areas.In contrast, plane 2, and no cells were visible on plane 4 ( Additionally, the experimental setup was disassembled perform WST-1 assays.On both occasions metabolic activity was observed to information on the synthesis of the scaffold was analyzed after 4 and pro-collagen increased from 304 to 355 ng/mL center (Figure 5b).To analyze the viability of human osteoblasts within the scaffold, cells were seeded on superior plane 1.One day after seeding, a lot of cells were observed on plane detected on planes 2 and 4.After 8 days of cultivation, a lot of viable cells 3.These cells also formed numerous cell connections both planes.It was also striking that a lot of dead cells could be determined within unpopulated areas.In contrast, there were only a few cells on the intermediate no cells were visible on plane 4 (Figure 4).

Mechanical properties of the three scaffold designs, given as means
Viability of human osteoblasts on the various planes of the 3D Ti6Al4V scaffold -h) of cultivation under static culture conditions (n dead cells = red; scale bar = 500 µm; (a), (e): plane 1; (b) : plane 4).
Additionally, the experimental setup was disassembled after days 1 and 8 of cultivation, in order occasions, metabolically-active cells were determined, whereby the was observed to increase two-fold during the incubation time (Figure 5a).To the ECM components, medium from the center and 8 days of cultivation, during which time the synthesis rate of type I to 355 ng/mL at the periphery, and from 255 to 391 ng/mL To analyze the viability of human osteoblasts within the scaffold, cells were seeded onto the superior plane 1.One day after seeding, a lot of cells were observed on planes 1 and 3, whereas no days of cultivation, a lot of viable cells were numerous cell connections, which resulted in a both planes.It was also striking that a lot of dead cells could be few cells on the intermediate planes of the 3D Ti6Al4V scaffold ) of cultivation under static culture conditions (n ≥ 3; living ), (f): plane 2; (c), after days 1 and 8 of cultivation, in order to determined, whereby the fold during the incubation time (Figure 5a).To obtain the center and periphery of the he synthesis rate of type I and from 255 to 391 ng/mL in the

Discussion
The challenge for a metallic implants adaptation of mechanical properties to the stress shielding [10,11,32].Adaptation to the mechanical properties of the bone must be considered, especially when dealing with open penetrate into the scaffold [33].
The use of additive manufacturing processes possibility to fabricate open-porous bone scaffolds shapes and structures.Furthermore, by gaining full control any desired mechanical properties can be directly Nevertheless, the mechanical properties of additive manufactured samples are similar to those manufactured conventionally.Koike samples fabricated by selective laser and electron beam melting results showed comparable yield and tensile strength melted and wrought specimens had slightly less strength.Furthermore, In the present study, the mechanical properties of open designs, were determined.Porosity c three scaffold designs.Although geometric structure and designs, ultimate compressive strength 164 MPa.These values are within the range determined by experimental testing In contrast, greater differences were observed between the three designs which varied between 3.7 and 6.7 properties of scaffolds may not only be influenced by porosity.activity of human osteoblasts seeded onto days of cultivation (n = 3); and (b) synthesis of pro-collagen type I human osteoblasts seeded onto the 3D Ti6Al4V scaffold.Supernatants were collected ultivation and analyzed using ELISA (n = 3).Data are .
metallic implants is to provide sufficient mechanical stability adaptation of mechanical properties to the surrounding tissue, in order to prevent bone loss due to Adaptation to the mechanical properties of the bone must be considered, open-porous scaffolds or surface coatings, where The use of additive manufacturing processes, such as laser or electron beam melting porous bone scaffolds, as well as non-porous implants res.Furthermore, by gaining full control of the shape and geometric composition, any desired mechanical properties can be directly controlled by the structural design.
Nevertheless, the mechanical properties of additive manufactured samples are similar to those conventionally.Koike et al. [17] compared the mechanical properties of tensile test samples fabricated by selective laser and electron beam melting with cast and wrought samples.results showed comparable yield and tensile strength values for the four different samples.Laser beam had similar strengths, whereas electron beam melted specimens .Furthermore, the yield strength of the cast samples was lowest the mechanical properties of open-porous scaffolds Porosity calculations revealed similar values, of approximate three scaffold designs.Although geometric structure and strut orientation differe designs, ultimate compressive strengths were similar for all scaffolds, varying MPa.These values are within the range of 140-220 MPa reported for human cortical bone, experimental testing [34,35].were observed between the three designs in terms of .7 and 6.7 GPa.These findings support the argument that the mechanical properties of scaffolds may not only be influenced by porosity.Since only three different structural the 3D Ti6Al4V collagen type I from scaffold.Supernatants were collected after Data are given as sufficient mechanical stability, and also in order to prevent bone loss due to Adaptation to the mechanical properties of the bone must be considered, the bone is supposed to laser or electron beam melting, offers the porous implants, in a wide range of the shape and geometric composition, by the structural design.Nevertheless, the mechanical properties of additive manufactured samples are similar to those compared the mechanical properties of tensile test with cast and wrought samples.Their four different samples.Laser beam whereas electron beam melted specimens had the cast samples was lowest of all.porous scaffolds, with three different of approximately 70%, for all orientation differed between the three varying between 145 and for human cortical bone, terms of elastic modulus, a.These findings support the argument that the mechanical three different structural designs were examined in the present study, our conclusions may not be significant, but nevertheless support previous findings published in the literature.
Murr et al. [20] fabricated cubic scaffolds of varying density using the electron beam melting (EBM) process.They found different mechanical property values for similar porosity values (82%).Scaffolds with strut thicknesses of 1.0 and 1.2 mm exhibited elastic moduli of 1.5 and 0.9 GPa, respectively.
Parthasarathy et al. [24] also used EBM to fabricate open-porous scaffolds with porosities ranging from 50% to 70%.All scaffold struts exhibited a rectangular cross-section, with two different strut thicknesses (450 and 800 µm).In general, the scaffolds demonstrated a decrease in mechanical properties with increasing porosity.In contrast, significantly different mechanical property values were observed, despite the fact that two of the scaffolds had similar porosities (approximately 50%).The resultant compressive stiffnesses were 2.9 and 0.6 GPa for scaffolds with strut thicknesses of 800 and 450 µm, respectively.
The elastic modulus results of Murr et al. [20] (2.7 GPa for 72% porosity) and Parthasarathy et al. [24] (2.1 GPa for 70% porosity) were lower than the values obtained during the current study (5.1 GPa for 70% porosity).These differences can be explained by variations in the fabrication methods (i.e., laser and electron beam melting).Samples fabricated by EBM were generally characterized by rougher surfaces than those made using SLM [17], and thus exhibited lower mechanical properties [36].Nevertheless, all elastic modulus results were lower than the values for human cortical bone, which fall in the range 15-20 GPa [34,35].
Regarding the biocompatibility of the examined material in vitro, the results of the current study suggest that human osteoblasts could survive on porous titanium scaffolds in a static cell culture.Furthermore, the synthesis of pro-collagen type I was not only sustained during the incubation period, it clearly increased.The scaffold macropores were settled by cells, although the pore size prevented an overgrowing of cells.Other studies also indicated the proliferation of human osteoblasts on scaffolds made of titanium [9,28].
Hollander et al. [28] seeded human osteoblasts onto the surface of additive manufactured titanium scaffolds with a regular, circular pore diameter of approximately 500-1000 µm.Their results showed proliferation and survival of the cells after 14 days of cultivation.Furthermore, the pores of the scaffolds filled with cells, which had grown along the pore rims.
Mueller et al. [9] seeded human osteoblasts onto foam-like open-porous titanium scaffolds with pore diameters between 100 and 750 µm.Under static culture conditions, and in a perfused system, they demonstrated that human osteoblasts grew through the interconnected pores of the metal foam, and expressed an osteoblast-like phenotype.
Cell proliferation was also observed during the present study, which implies bone ingrowth into titanium implants in a biological environment, as described by Mangano et al. [23].
It should be noted that the cell investigations presented here were performed on only one geometric scaffold type.It is possible that cell behavior could vary between pores of different geometric shape.Nevertheless, methodical investigations of cell proliferation in different pore sizes are rare.
Frosch et al. [37] determined the effect of different diameters of cylindrical titanium channels on human osteoblast cell proliferation.Pores with diameters of between 300 and 1000 µm were drilled into a titanium block, such that there were no interconnections.The experiments indicated that channel diameter had no influence on collagen type I production.Furthermore, the highest differentiation was found in 600 µm pores, whereas the highest cell density was in 300 µm pores.
In summary, the current study demonstrated the influence of three different scaffold designs on mechanical properties, providing an open-porous design with adequate pore geometry [9].Furthermore, it was shown that a low elastic modulus can stimulate new bone formation, due to mechanical stimulus by physiological load application, and avoid stress shielding caused by high stiffness gradients between bones and implants [10,11].Nevertheless, in instances of large segmental defects in long bones, initial stabilization of open-porous scaffolds should be supported by osteosynthesis systems, such as intramedullary nails, plates or by external fixation.Consequently, stress distribution within the bone-implant interface in such a complex situation should be further analyzed, in order to obtain a realistic prediction of the situation in vivo.The results of the in vitro testing indicated a high degree of human osteoblast (which are very sensitive to artificial materials) cell proliferation on the titanium surface.

Conclusions
Using additive manufacturing process SLM, the design diversity and adaptation to nearly any desired target value can be implemented to open-porous bone scaffolds.Furthermore, the mechanical load on the bone can be controlled and consequently the stimulation for bone regeneration as well as stress shielding due to the optimized material properties.It is assumed that the proliferation and survival of cells onto the titanium surface would lead to complete bone ingrowth into the implant under in vivo conditions.

Figure 1 .
Figure 1.CAD models of the three investigated structures struts aligned in the vertical direction and (c) scaffold with diagonally

Figure 2 .
Figure 2. (a) Scaffold with identical strut design in shifted strut orientation; and 1339(b) scaffold with and also for the additive were calculated according to the following equations:(1)is the overall volume enclosed by the(2) is the density of the manufactured mechanical properties.Axial carried out using a universal testing machine (Z50; mm/min for all scaffolds.Values of applied load and displacement of manufactured test samples,

Figure 3 .
Figure 3. Engineering stress scaffold.One graph is plotted for each type of scaffold struts; (2) rectangular struts struts.Photographs of the mechanical failure for each type of scaffold the right.

Figure 5 .
Figure 5. (a) Metabolic activity of human osteoblasts seeded on scaffold, after 8 days of cultivation (n human osteoblasts seeded on 4 and 8 days of cultivation and analyzed means ± standard deviations.

Table 1 .
Porosities of the three scaffold types, calculated from the CAD and for the manufactured samples (AMS), given as means ± (in the case of AMS) standard deviations.

Table 2 .
Mechanical properties of the three scaffold designs