- Email: [email protected]
Novel multilayer Ti foam with cortical bone strength and cytocompatibility
Acta Biomaterialia 9 (2013) 5802–5809
Contents lists available at SciVerse ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Novel multilayer Ti foam with cortical bone strength and cytocompatibility K. Kato a,⇑, S. Ochiai b, A. Yamamoto c,d, Y. Daigo e, K. Honma e, S. Matano b, K. Omori c a
Mitsubishi Material Corp., 1-297 Kitabukuro-cho, Omiya-ku, Saitama 330-8508, Japan Department of Materials Science and Engineering, Graduate School of Engineering, Kyoto University, Yoshida, Sakyo-ku, Kyoto 606-8501, Japan c Biomaterials Center, National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan d International Center for Materials Nanoarchitectonics (WPI-MANA), National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan e Mitsubishi Materials Corp., Advanced & Tools Company, High Performance Alloy Products Divison, 476 Shimoishido-shimo, Kitamoto, Saitama 364-0023, Japan b
a r t i c l e
i n f o
Article history: Received 24 August 2012 Received in revised form 31 October 2012 Accepted 13 November 2012 Available online 29 November 2012 Keywords: Porous Ti Ti foam Cortical bone Stress shielding Cytocompatibility
a b s t r a c t The major functions required for load-bearing orthopaedic implants are load-bearing and mechanical or biological fixation with the surrounding bone. Porous materials with appropriate mechanical properties and adequate pore structure for fixation are promising candidates for load-bearing implant material. In previous work, the authors developed a novel titanium (Ti) foam sheet 1–2 mm thick by an original slurry foaming method. In the present work, novel Ti foam is developed with mechanical properties compatible with cortical bone and biological fixation capabilities by layer-by-layer stacking of different foam sheets with volumetric porosities of 80% and 17%. The resulting multilayer Ti foam exhibited a Young’s modulus of 11–12 GPa and yield strength of 150–240 MPa in compression tests. In vitro cell culture on the sample revealed good cell penetration in the higher-porosity foam (80% volumetric porosity), which reached 1.2 mm for 21 days of incubation. Cell penetration into the high-porosity layers of a multilayer sample was good and not influenced by the lower-porosity layers. Calcification was also observed in the highporosity foam, suggesting that this Ti foam does not inhibit bone formation. Contradictory requirements for high volumetric porosity and high strength were attained by role-sharing between the foam sheets of different porosities. The unique characteristics of the present multilayer Ti foam make them attractive for application in the field of orthopaedics. Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction As the average age of the population is increasing in the developed world, the demand for load-bearing metals for orthopaedic implants has grown drastically [1]. The major functions required of orthopaedic implants are load bearing and mechanical or biological fixation with the surrounding bone. Ti and its alloys have been used in orthopaedic implants for load-bearing applications because of their biocompatibility, good corrosion resistance and excellent mechanical properties. However, the major problem of Ti and Ti alloy implants is a mismatch of Young’s modulus between bone (10– 30 GPa) and Ti (110 GPa), which leads to stress-shielding and results in bone resorption and implant loosening [2]. One possible solution to this problem is the development of low-modulus Ti alloys, which, over the past 20 years, has resulted in new alloys such as TNTZ (Ti–29Nb–13Ta–4.6Zr) [3,4]. Unfortunately, the Young’s modulus of human bone is too low to be reached by an alloydesigning approach. An alternative method of overcoming the mismatch of Young’s modulus is the use of porous metal coating, ⇑ Corresponding author. Tel.: +81 48 641 5111; fax: +81 48 641 6313. E-mail address: [email protected] (K. Kato).
which can reduce mismatch at the interface and achieve stable long-term biological fixation through bone ingrowths. Many types of porous metals have been developed and used to coat the surface of orthopaedic implants [5–10]. Typical fabrication methods for such porous metals are powder metallurgy [11–13], chemical vapour infiltration [14,15], space-holding [16,17], plasma spraying [18] and rapid prototyping [19–21]. Although open-pore structures have the advantages of a low Young’s modulus and bone ingrowth, leading to better fixation with bone, they have the disadvantage of insufficient mechanical strength compared with that of bulk structural materials [22]. As a result, the majority of these porous metals are applied as coatings on fully dense substrates [9]. Imwinkelried [16] reported the first application of a Ti foam device for the human lumbar spine. However, the yield strengths of this Ti foam, which had a porosity of 63–50%, were 68–140 MPa under compression. This low mechanical property limited its application to the load-bearing devices for the human lumbar spine. Therefore, porous metal fabrication technology ensuring adequate strength and tailorable open-pore structure is assumed to be very important for orthopaedic implants. To achieve these contradictory requirements of adequate strength and high porosity with adequately open-pore structure, the present authors developed novel
1742-7061/$ - see front matter Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actbio.2012.11.018
5803
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
Ti foam by stacking foam sheets of different porosities layer by layer to create multilayer Ti foam. This study examines the mechanical properties of multilayer Ti foam designed to have cortical bone strength. Cell penetration capability is evaluated by in vitro experiments and compared with a control sample prepared by stacking a single type of foam sheet. In addition, cell penetration and calcification in the Ti foam sheet are examined. 2. Materials and methods 2.1. Manufacturing process and specification of Ti foam and multilayer Ti foam An original slurry foaming method, illustrated in Fig. 1a, was used to fabricate novel Ti foam. A slurry containing Ti powder (20 lm in mean diameter), binder, foaming agent and water was prepared and coated on a carrier sheet. The sheet was treated in a foaming moisture chamber, and then moved to a drying chamber, where the foaming agent in the slurry was evaporated. Foamed green sheets were heated for debinding and then sintered in a vacuum heat-treatment furnace. This Ti foam has already been used commercially in new electrochemical applications [23]. As-fabricated Ti foam sheets were 0.2–2 mm thick and possessed a prescribed porosity and pore diameter. The volumetric porosity of the foam P was estimated by
P ¼ 1 ðq =qs Þ where q⁄ is the measured apparent density of foam, and qs is the density of bulk Ti. Because the Ti foam had a sheet shape, it was possible to stack them easily, despite their different volumetric porosities. Table 1 lists the test samples prepared by stacking the Ti foam sheets. Designation letters indicate Ti foam with high (H), middle (M) or low (L) volumetric porosity. Designation numbers indicate the approximate thickness in millimetres of the corresponding Ti foam in the stacked sheet. The ‘‘multilayer’’ type consists of two different volumetric porosity sheets, and the ‘‘stacked’’ type consists of the same volumetric porosity sheets. Table 2 lists the chemical composition of Ti powder and Ti foam. A photograph and microscopic images of H2L1, multilayer Ti foam composed of high volumetric porosity sheet (H, 80%) and low volumetric porosity sheet (L, 17%), are shown in Fig. 1b–d. As specified in Table 1, H2L1 consisted of four pairs of eight sheets: high volumetric porosity sheet with a thickness of 2 mm and low volumetric porosity sheet with a thickness of 1 mm alternated four times. The sample was sintered in a vacuum heat-treatment furnace. In the case of M2L2, M denotes a middle volumetric porosity sheet of 62%, and 2 denotes a thickness of 2 mm. Similarly, H2 was composed of a single type of 2-mm-thick high volumetric porosity Ti foam sheet. H2 data were obtained during a previous work [23].
(a) Slurry ; Ti Powder, Binder, Foaming Agent Blade Ti Foam
Carrier Sheet
Mixing
Coating
(b)
Foaming
Drying
(c)
100µm
Debinding
Sintering
(d)
10µm
Fig. 1. (a) Illustration of the slurry foaming method; (b) photograph of a multilayer Ti foam sample; (c) and (d) scanning electron microscopy images of the arrow surface of (b) in (c) low and (d) high magnification.
Table 1 Specifications of multilayer and stacked Ti foam samples. Designation
H1M1 H2L1 H1L1 M2L2 H2e M2
Type
Multilayer Multilayer Multilayer Multilayer Stacked Stacked
Total porositya (%)
72 57 42 38 80 62
No. of repetitions
6 4 6 3 6 6
Ti foam Ab
Ti foam Bb
DA (lm)
PA (%)
TA (mm)
DB (lm)
PB (%)
TB (mm)
320 320 380 340 300 340
80 80 80 62 80 62
1 2 1 2 2 2
320 –c –c –c – –
63 17 17 17 – –
1 1 1 2d – –
D, average pore diameter; P, porosity; T, thickness. Designation letters indicate Ti foam with high (H), middle (M) or low (L) volumetric porosity. Designation numbers indicate the approximate thickness of the corresponding Ti foam in the stacked sheet. a Total porosity of multilayer and stacked foam was slightly lower than the initial porosity of Ti foam A and B, because of the stacking sintering process. b Properties of Ti foam A and foam B are measured in the sheet condition. c Pore diameter of the foam with low porosity was difficult to measure by optical microscopy. d Ti foam L was stacked using two sheets of L1. e Data of stacked sample H2 was obtained in previous work [23].
5804
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
Table 2 Chemical composition of Ti powder and foam (mass%).
Ti powdera Ti foam Ab Ti foam Bc
c
H
Fe
O
N
<0.01 0.10 0.13
0.034 0.008 <0.001
0.065 0.054 0.040
0.23 0.27 0.30
0.01 0.01 0.01
Mill certificate. Chemical analyses in previous work [23]. Chemical analyses in present work.
Stacked thickness Stacked thickness
Length
a b
(b)
(a)
C
a
Designation
Average pore diametera (lm)
Porosity (%)
Dimensions (mm)
Type
Assay
H180
180
87
W12, L10, T2
Penetration
H320
320
85
W12, L10, T2
H500
500
85
W12, L10, T2
H150
150
89
W11, L11, T0.5
Single sheet Single sheet Single sheet Single sheet
Penetration
Length
Table 3 Specifications of Ti foam sheet samples.
(c) Width
Penetration Calcification
Average pore diameter was measured by digital image processing.
The pore diameter of the Ti foam was controllable between 50 and 500 lm. In the present work, four types of Ti foam sheets with different pore diameters for cell penetration and calcification assays were prepared, as shown in Table 3.
Thickness TB
TA
Fig. 2. Illustration of compression and tension test samples. Arrows show the compression and tension directions. (a) Stacked Ti foam sample; (b) multilayer Ti foam sample; (c) single Ti foam sheet. TA and TB represent the thickness of Ti foam sheets A and B in the multilayer foam.
2.2. Mechanical test samples and test procedures Mechanical test samples are listed in Table 1. Three samples of each type were tested. The compression test samples were 12 mm wide, 10 mm long and 11–12 mm in stacked thickness, as shown in Fig. 2a. The number of repetitions of foam sheet pairs was chosen to give nearly the same stacking thickness for all samples of 12 mm. The single sheet compression test sample L was 5 mm wide, 4.5 mm long and 1 mm thick, as illustrated in Fig. 2c. Compression tests were carried out at room temperature with a tensile and compression testing machine (TG-20kN, Minebea Corp, Nagano, Japan) according to JIS H 7902: Method for compressive test of porous metals. Testing was conducted at a constant crosshead speed of 0.5 mm min1. Strain measurements were carried out using two contact digital displacement sensors (AT-010, Keyence Corp, Osaka, Japan), with which the displacement of the sandwich board was accurately determined. The compression and tensile test directions were parallel to the stacking interface. The tensile test samples were 5 mm wide, 50 mm long and 11–12 mm in stacked thickness, in the directions shown in Fig. 2b. Both ends of the tensile samples were glued to aluminium tabs with epoxy-glue [23]. Tensile testing was carried out at room temperature with a tensile machine (AUTOGRAPH-AG 50kNG, Shimadzu Corp., Kyoto, Japan) at a constant crosshead speed of 8.3 106 m s1. Sample strain was measured with a non-contact extensometer [24–26]. Two paper markers were pasted onto the samples at a spacing of 20 mm, and their displacement was recorded with a laser CCD camera (DVW-200 Shimadzu Corp.). 2.3. Cell penetration assay of Ti foam sheet, stacked and multilayer Ti foam First, the effect of pore diameter on cell penetration was evaluated using the foam sheet samples (see Table 3) and, second, cell
penetration into stacked foam H2 and multilayer foam H2L1 was evaluated (see Table 1 and Fig. 3a). The stacked and multilayer samples were 2 mm wide, 10 mm long and 11–12 mm in stacked thickness. Cell penetration assays were performed in the manner shown schematically in Fig. 3b. In most cytocompatibility tests for porous materials, cell suspension is directly applied onto the porous samples. As a result, some of the cells stay at the very top surface and some drop through the pores, and thus inoculation itself causes some difference in cell penetration distance at the very first stage of cell culture. To avoid such error in this study, human osteosarcoma Saos-2 cells were seeded on a sterile silicone sheet (15 mm square and 0.1 mm thick) at a concentration of 1.85 105 cells per 6 ml of Dulbecco’s modified minimum essential medium supplemented with 10 vol.% foetal bovine serum (D-MEM + FBS), cultured for 4 days to be confluent, and the resulting cell-covered sheet was then attached to the autoclaved specimen surface, as shown in Fig. 3b. A glass ring 9 mm in diameter was placed over the silicone sheet to prevent movement of the sheet. The cells were cultured for 7, 14 and 21 days, while the culture medium was replaced with fresh medium every other day. Then, the cells were fixed with a mixture of 10 vol.% formalin and 10 vol.% methanol. The sample was sliced longitudinally with a diamond cutter, as shown in Fig. 3b, and the cells were stained with sulforhodamine 101 acid chloride. Reproducibility was confirmed by repeating the same experiment. Cells in the sample were observed from the side (longitudinal plane) with a fluorescent microscope (MVX-10, Olympus Co, Ltd., Tokyo, Japan). The farthest distance from the specimen surface to cells in a view field of 3.52 2.68 mm was measured and averaged for more than four view fields per sample. The statistical significance of data obtained for three different single sheets was analysed by the Tukey (or Tukey–Kramer) test, while those for
5805
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
(a) 1mm
Subsequently, the culture medium was replaced with calcification medium; D-MEM + FBS containing 0.5 mM b-glycerophosphate and 50 lg ml1 L-ascorbic acid. The cells were cultured for 7, 14, 21 and 28 days, while the medium was exchanged every other day. Calcein was then added to give a concentration of 1 lg ml1, and the microplate was kept in the CO2 incubator for a further 4 h. Then, the cells were fixed and stained in the manner described above. The cells in the sample were observed by a confocal laser scanning microscope (LSM-510, Carl Zeiss Japan, Tokyo, Japan) for more than three views for one sample. The reproducibility was confirmed by repeating the same experiment. 3. Results 3.1. Mechanical properties of the multilayer Ti foam
(b)
cell
4d of incubation
metal foam 7, 14, or 21d of incubation
Fixation Cutting Fluorescent staining
Observation by fluorescent microscopy
The compression and tensile test results are listed in Table 4. In compression tests, the Young’s modulus and yield strength of multilayer samples with L1 or L2 were 11 GPa and 150–250 MPa, respectively. It is notable that these results for the present multilayer Ti foam with L1 or L2 are comparable with those for cortical bone, the Young’s modulus and strength of which under compression are 10–30 GPa and 150–220 MPa [27–29], respectively. In tensile tests, the Young’s modulus and strength of multilayer samples with L1 or L2 were 17–22 GPa and 100–110 MPa, respectively. These results are also comparable with those of cortical bone, the Young’s modulus and strength of which are 10–25 GPa and 90–170 MPa, respectively [27–29]. Tensile Young’s moduli were higher than the compression moduli, because of uniform stress distribution in the foam. In tensile tests, the grip section of the sample was glued to avoid stress concentration and, furthermore, stress–strain curves were measured in the gauge section
Cell Penetration Distance (mm)
silicone sheet
Fig. 3. (a) Photographs of multilayer Ti foam sample H2L1 for cell penetration assay and (b) schematic explanation of cell penetration assay procedure.
stacked and multilayer foam were analysed by an unpaired t-test (or Welch’s modified t-test). 2.4. Calcein assay of Ti foam sheet A foam sheet sample 11 mm square and 0.5 mm thick was set into a silicone O-ring to prevent its direct contact with the bottom of the culture vessel surface to avoid penetration of cells growing on the culture vessel surface into the foam sheet. The foam sheet in the silicone O-ring was autoclaved and then placed on the well of a ‘‘non-treated’’ 12-well microplate. Saos-2 was inoculated onto the sample at a concentration of 1 106 cells in 0.1 ml DMEM + FBS. After 20 min, a 2 ml portion of D-MEM + FBS was added, and the cells were then cultured for 3 days to be confluent.
1.5
H180 H320 H500
1.0
0.5
*
mean s.d.
0 0
5
10
15
20
25
Incubation Period (d) Fig. 4. Cell penetration into Ti foam sheets with different open-pore sizes. Subscripts in graph indicate mean pore diameter (see Table 2). ⁄Significant difference is only observed between H320 and H500 after 7 days’ incubation (p < 0.05).
Table 4 Compresion and tensile test results for multilayer and stacked Ti foam samples. Designation
Porosity (%)
Compression yield strength (MPa)
Compression Young’s modulus (GPa)
Tensile strength (MPa)
Tensile Young’s modulus (GPa)
Nominal tensile fracture strain of tensile test (%)
H1M1 H2L1 H1L1 M2L2 H2a M2 Lb
72 57 42 38 80 62 17
34.3 ± 0.6 158.3 ± 14.6 240.0 ± 10.0 251.7 ± 2.9 19.4 ± 4.1 47.8 ± 0.2 431.0 ± 28.2
2.8 ± 0.2 11.1 ± 2.3 11.5 ± 1.4 11.1 ± 0.5 3.2 ± 0.6 6.4 ± 0.2 17.1 ± 1.1
25.5 ± 1.8 110.1 ± 19.3 103.7 ± 4.9 113.8 ± 6.5 20.2 ± 2.1 – –
5.8 ± 1.2 21.7 ± 6.7 17.6 ± 1.2 19.1 ± 2.7 3.8 ± 1.2 – –
0.98 ± 0.31 0.72 ± 0.15 0.68 ± 0.17 0.58 ± 0.05 1.35 ± 0.86 – –
Test results are mean ± standard deviation. Designation letters indicate Ti foam with high (H), middle (M) or low (L) volumetric porosity. Designation numbers indicate the approximate thickness of the corresponding Ti foam in the stacked sheet. a Obtained in previous work [23]. b Single Ti foam sheet. Compression was loaded in the direction of the sheet thickness, assuming isotropy of the foam sheet L.
5806
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
Stacked foam (H2)
Multilayer Ti foam (H2L1) cell contacting surface
7d
0.5mm
0.5mm
0.5mm
0.5mm
0.5mm
0.5mm
14d
21d
0.5mm
0.5mm
Cell Penetration Distance (mm)
Fig. 5. Fluorescence microscopy images of cells penetrating samples H2 and H2L1. Arrows indicate furthest cell penetration distance in the imaged area. The bottom images show longitudinal cross-sections of both samples.
1.5 H2 H2L1 1.0
0.5 mean s.d. 0
0
5
10
15
20
25
Incubation Period (d) Fig. 6. Cell penetration into samples H2 and H2L1. No significant difference was observed between H2 and H2L1 for all incubation periods.
apart from the grapping portion. Because of this specially designed test method, tensile stress was distributed uniformly in cross section. In the compression test, the sample foam surface contact to the compression tool plate was relatively uneven, which did not give uniform stress in cross section. In contrast to the compression
yield strength, the tensile strength did not increase with decreasing porosity, as shown in Table 4. This difference may be attributed to the lack of uniform pore dispersion; in tensile testing, any irregular porosity brings about stress concentration, which causes early failure. In Fig. 8, the compression yield strength data in Tables 4 and 5 are plotted against the porosity. Two models were applied to these results to evaluate the effect of stacking low-porosity and highporosity layers on the yield strength of the multilayer foam: the Gibson–Ashby model [30] (Eq. (4.1) shown later in Section 4.1) and the rule of mixtures (Eq. (4.2) shown later in Section 4.1). The curved line calculated by Eq. (4.1) and the straight lines calculated by Eq. (4.2) are also shown in Fig. 8. 3.2. Cell penetration in the Ti foam sheet, stacked and multilayer Ti foam Fig. 4 presents the results of the cell penetration assay for the three types of Ti foam sheet with different pore diameters. In comparison with H180, H320 had larger cell penetration distances for all
5807
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
7d
alongside the cells, suggesting the importance of cell penetration into the Ti foam.
14d
4. Discussion 4.1. Mechanical properties
100µm
100µm
21d
28d
100µm
100µm
28d
Most mechanical tests of Ti foam in the literature have been by compression testing. Therefore, the compression properties of H1L1, produced by the slurry foaming method, were compared with those of samples produced by other methods, with volumetric porosities ranging from 33% to 50%, as shown in Table 5. H1L1 had a superior 0.2% yield strength compared with the other samples. The reason for this stems from the existence of the 17% porosity layer, which had a high yield strength of 431 MPa, as measured in the present work (see Table 4). For reference, the 0.2% yield strength of the stacked sample of 80% porosity sheets was 19.4 MPa, estimated in recent work [23] (also shown in Table 4). In Fig. 8, the concrete form of the equation for each model is summarized as follows, with important indications obtained by comparison of the experimental data with the calculations. According to Gibson–Ashby [30], the yield strength (rpl ) of open-pore foam under compression can be expressed as a function of the relative density (q⁄/qs), with the apparent densities of the foam q⁄ and the solid material qs.
rpl =rys ¼ C ðq =qs Þ3=2
ð4:1Þ
In Eq. (4.1), rys and C are the yield strength of the solid material and a constant of proportionality, respectively. Substituting the yield strength rys = 650 MPa [16] and the test results obtained for H2 in into Eq. (4.1), one has C = 0.33, which is in good agreement with the 0.3 used in Gibson–Ashby’s work [30]. However, one can draw straight lines between L and H2 and between L and M2. These lines can be expressed using the volumetric ratio method given by Eq. (4.2), in which the yield strength rm of multilayer foam is approximated to be given by the rule of mixtures for the yield strength values of the constituting foam sheets.
100µm
rm ¼ rA TA =ðTA þ T B Þ þ rB TB =ðTA þ T B Þ
Fig. 7. Calcification of Saos-2 on a single Ti foam sheet, H150. Green indicates densely calcified areas marked by calcein, and orange indicates cells marked mainly by Texas Red (and slightly by calcein as well).
incubation periods, reaching 1.2 mm after 21 days’ incubation. In contrast, H500 showed a cell penetration distance similar to that of H320 for 7 and 14 days’ incubation, but remained 1.0 mm for 21 days’ incubation. Results for the stacked and multilayer samples, H2 and H2L1, are shown in Figs. 5 and 6. Cell penetration distance tended to increase with longer incubation period, and no difference between H2 and H2L1 was observed. This confirmed that maximum cell penetration into the multilayer sample was obtained at the higher porosity layers and was not influenced by the lower porosity layers.
3.3. Calcification in the Ti foam sheet The results of the calcification assay for H150 are shown in Fig. 7, in which calcification of Saos-2 in the Ti foam was observed. The densely calcified area clearly increased with incubation period, and was especially apparent after 28 days’ culture. The magnified images at the bottom of Fig. 7 revealed that calcification occurred
ð4:2Þ
Here, r and r are the yield strengths, and TA and TB the thicknesses of foam A and foam B, respectively. The following conclusions are read from the experimental data and calculation results shown in Fig. 8. (1) The curved line of Eq. (4.1) approximately describes the experimental data obtained for samples with relatively low yield strength; the single or stacked samples and multilayer samples composed of relatively high-porosity foam. In contrast, the volumetric ratio method ((4.2)) approximately describes the data obtained for the multilayer samples. This finding suggests that the mechanism controlling the yield strength of multilayer samples with closed pore foam was different from that of those with uni A
B
Table 5 Compression test results for Ti foam produced by slurry foaming, rapid prototyping, space holder, and plasma-spray methods. Methods
Porosity (%)
Pore diameter (lm)
Yield strength (MPa)
Young’s modulus (GPa)
Ref.
Slurry foaming, multilayer Ti foam Rapid prototyping Space holder Plasma spray
42
380
230–250
10–13
Present work
33–40
700
100–210
13–20
[20]
50 40
100–500 300–500
130–140 85
16–18 4–6
[16] [18]
5808
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
500
450 0.2% Yield Strength, MPa
Present work with porosity 17% layers Present work without porosity 17% layers Porosity 17% sheet Previous work [23] Ref erence data
L sheet
400
350 300 250
H1L1
M2L2
200
[20]
H2L1
150
Gibson-Ashby 100 Equation curve
[16] [18]
50
H2
M2 H1M1
0 0
10
20
30
40
50
60
70
80
90
Porosity, % Fig. 8. Compression 0.2% yield strength plotted against porosity.
form open-pore foam. (2) The straight line ((4.2)) for the foam containing 17% porosity L sheets exhibited a higher yield strength than the curved line of Eq. (4.1) for open-pore foam. This means that the multilayer Ti foam had a yield strength superior to porous materials, whose yield strength followed Eq. (4.1). (3) The 0.2% yield strength of the present multilayer Ti foam is thus controllable by changing the volumetric ratio of higher-porosity foam A to lower-porosity foam B, as shown in Table 1. 4.2. Cell penetration In vivo implantation tests into animal bone tissue are generally performed to evaluate the biological fixation ability of porous materials. Such testing takes a long period of time, is costly and has limitations on the number of implanting samples. Most importantly, differences in species between animals and humans often bring different results. Therefore, animal experiments are expected to be refined, reduced and replaced by other evaluation methods such as in vitro experiments. For bone formation, either body fluid or cells must penetrate into the foam, which is influenced by the foam structure. Using appropriate experimental conditions, it is possible to evaluate cell penetration into the foam by in vitro experiments as a screening test. However, studies on cytocompatibility of porous Ti or other metals with different pore structures have only evaluated cell proliferation. No research has been performed as yet to evaluate cell penetration into metal foam. Therefore, cell penetration was evaluated for different pore diameters. It was found that H320 had the most stable and largest cell penetration distance among the foam sheets tested. This may be because of a balance between the opportunity for cells to transfer from the silicone sheet to the pore walls and penetration efficiency for deeper penetration into the foam. A larger pore diameter can offer higher efficiency to gain penetration distance, but also less opportunity for cells to transfer to the foam. H320 was considered to have an optimal balance between penetration efficiency and transferral opportunity among the foam sheets tested in the cell penetration assay. Based on in vivo experiments, 100 lm is generally considered the minimum recommended pore size for bone ingrowth, but pores >300 lm have better bone formation, because relatively
larger pores favour direct osteogenesis because of their higher vascularization and oxygenation [31]. Pore sizes >1 mm, however, increase fibrous tissue formation [32]. The pore size of H320 agreed with the favourable range described above. Therefore, foam of this diameter was employed in the preparation of stacked and multilayer samples to support cell penetration. Cell penetration assays revealed that multilayer samples including this foam were able to maintain good cell penetration in the higher-porosity layers with no influence of the lower-porosity layers. Combining Ti foam sheets with different porosities or pore structures while maintaining relatively high cell penetration increases the range of orthopaedic load-bearing device designs. The present multilayer Ti foam, with a combination of lower- and higher-porosity foam sheets, exceeds the limitations of previous foam materials by sharing the contradictory roles of strength and cell penetration.
4.3. Cell calcification Calcification is one of the steps occurring at the early stage of bone formation in vivo, and is a good measure of bone tissue mineralization. Although in vitro calcification assay is often performed to examine cellular differentiation and maturity of osteogenic cells, it can also be used to evaluate the ability of materials to support the differentiation and mineralization of cells cultured on them. In the present study, calcification of Saos-2 in the single Ti foam sheet was confirmed, as shown in Fig. 7. Densely calcified areas were observed alongside the cells, suggesting the importance of cell penetration into Ti foam for bone formation. It has been reported that, for a Ti specimen with drilled channels of different diameters, the amount of mineralization of human osteoblasts in the channels correlates well with the cell penetration distance [33]. These results indicate the effectiveness of cell penetration assay as an in vitro evaluation method to estimate bone formation in porous materials. Thus, based on the observed good cell penetration and evidence of calcification, this Ti foam was expected to be suitable for implant applications requiring fixation with bone.
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
4.4. Application Spacers, such as interbody fusion cages for the human lumbar spine, are one typical orthopaedic load-bearing device. The current major materials used for these cages are solid Ti and Ti alloys, because of their high strength. However, failed fusion and cage malposition or migration have been identified in such devices [9]. Recently, Fujibayashi et al. [34] reported a novel porous Ti sintered with a solid outer frame for clinical use, which achieved successful spine fusion within 6 months. They explained the importance of spacer materials with high strength and high bone-bonding ability for spinal fusion without autologous iliac crest bone grafting [34]. The present multilayer Ti foam has adequate Young’s modulus, high yield strength in compression because it has low-porosity layers, and good cell penetration because it also has high-porosity layers. Furthermore, its high yield strength compared with other porous materials can be controlled by changing the thickness ratio of the high-porosity layer to low-porosity layer. These characteristics of multilayer Ti foam make it promising for load-bearing implant materials in the field of orthopaedic spinal devices. 5. Conclusions A novel multilayer Ti foam with mechanical properties compatible for cortical bone and biological fixation capability was developed. The Young’s modulus and yield strength of the foam were controllable to the range of those exhibited by cortical bone. Good cell penetration and calcification was also observed during in vitro screening tests. Cell penetration assay of a single Ti foam sheet showed that Ti foam with a 320 lm pore diameter and 85% volumetric porosity had optimum cell penetration distance among the three foam sheets tested in this study. A cell penetration assay for a multilayer sample containing this foam sheet revealed good cell penetration in the higher porosity layers without any influence of the lower porosity layers. Calcification of Saos-2 was also confirmed in the single Ti foam sheet. The multilayer Ti foam solves the contradictory requirements of adequate strength and high porosity with adequate open-pore structure. The unique characteristics of multilayer Ti foam make it an attractive load-bearing implant material, especially in the field of spinal orthopaedics. Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 1, 3, 5 and 7, are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/ j.actbio.2012.11.018. References [1] Nasab MB, Hassan MR, Sahari BB. Metallic biomaterials of knee and hip – a review. Trend Biomater Artif Organs 2010;24:69–82. [2] Krishna BV, Bose S, Bandyopadhyay A. Low stiffness porous Ti structures for load-bearing implants. Acta Biomater 2007;3:997–1006. [3] Akahori T, Niinomi M, Koyanagi Y, Kasuga T, Toda H, Fukui H, et al. Mechanical properties of biocompatible beta-type titanium alloy coated with calcium phosphate invert glass-ceramic layer. Mater Trans 2005;46:1564–9. [4] Tutumi H, Niinomi M, Akahori T, Nakai M, Takeuchi T, Katsura S. Mechanical properties of a beta-type titanium alloy cast using a calcia mold for biomedical applications. Mater Trans 2010;51:136–42. [5] Hahn H, Palich W. Preliminary evaluation of porous metal surfaced titanium for orthopedic implants. J Biomed Mater 1970;4:571–7. [6] Bobyn JD, Pilliar RM, Cameron HU, Weatherly GC. The optimum pore size for the fixation of porous- surfaced metal implants by the ingrowth of bone. Clin Orthop Relat Res 1980;150:263–70.
5809
[7] Levine B. A new era in porous metal: applications in orthopedics. Adv Eng Mater 2008;10:788–92. [8] Wazen RM, Lefebvre LP, Baril E, Nanci A. Initial evaluation of bone ingrowth into a novel porous titanium coating. J Biomed Mater Res B 2010;94B:64–71. [9] Singh R, Lee PD, Dashwood RJ, Lindley TC. Titanium foams for biomedical applications: a review. Mater Technol 2010;25:127–36. [10] Mour M, Das D, Winkler T, Hoenig E, Mielke G, Morlock MM, et al. Advances in porous biomaterials for dental and orthopedics application – review. Materials 2010;3:2947–74. [11] Oh IH, Nomura N, Masahashi N, Hanada S. Mechanical properties of porous titanium compacts prepared by powder sintering. Scripta Mater 2003;49:1197–202. [12] Chen J, Paetzell E, Zhou J, Lyons L, Soboyejo W. Osteoblast-like cell ingrowth, adhesion and proliferation on porous Ti–6Al–4V with particulate and fiber scaffolds. Mater Sci Eng 2010;C30:647–56. [13] Torres Y, Pavon JJ, Nieto I, Rodriguez JA. Conventional powder metallurgy process and characterization of porous titanium for biomedical applications. Metall Trans B 2011;42B:891–900. [14] Zardiackas LD, Parsell DE, Dillon LD, Mitchell DW, Nunnery LA, Poggie R. Structure, metallurgy, and mechanical properties of a porous tantalum foam. J Biomed Mater Res Appl Biomater 2001;58:180–7. [15] Shimko DA, Shimko VF, Sander ED, Dickson KF, Nauman EA. Effect of porosity on the fluid flow characteristics and mechanical properties of tantalum scaffolds. J Biomed Mater Res 2005;73:315–24. [16] Imwinkelried T. Mechanical properties of open-pore titanium pore. J Biomed Mater Res 2007;81A:964–70. [17] Hong TF, Guo ZX, Yang R. Fabrication of porous titanium scaffold materials by a fugitive filler method. J Mater Sci Mater Med 2008;19:3489–95. [18] Takemoto M, Fujibayashi S, Neo M, Suzuki J, Kokubo T, Nakamura T. Mechanical properties and osteoconductivity of porous bioactive titanium. Biomaterials 2005;26:6014–23. [19] L-Heredia MA, Sohier J, Gaillard C, Quillard S, Dorget M, Layrolle P. Rapid prototyped porous titanium coated with calcium phosphate as a scaffold for bone tissue engineering. Biomaterials 2008;29:2608–15. [20] Ryan GE, Pandit AS, Apatsidis DS. Porous titanium scaffolds fabricated using a rapid prototyping and powder metallurgy technique. Biomaterials 2008;29:3625–35. [21] Shishkovskii IV, Yadroitsev IA, Smurov. Selective laser sintering/melting of nitinol–hydroxyapatite composite for medical application. Powder Metall Met Ceram 2011;50:275–83. [22] Balla VK, Bodhak Z, Bose S, Bandyopadhaya A. Porous tantalum structures for bone implants: fabrication, mechanical and vitro biological properties. Acta Biomater 2010;6:3349–59. [23] Kato K, Yamamoto A, Ochiai S, Daigo D, Isobe T, Matano S, et al. Cell proliferation, corrosion resistance and mechanical properties of novel titanium foam with sheet shape. Mater Trans 2012;53:724–32. [24] Ochiai S, Nakano S, Fukazawa Y, Aly MS, Okuda H, Kato K, et al. Change of Young’s modulus with increasing applied tensile strain in open cell nickel and copper foams. Mater Trans 2010;51:925–32. [25] Ochiai S, Nakano S, Fukazawa Y, Aly MS, Okuda H, Kato K, et al. Tensile deformation and failure behavior of open cell nickel and copper foams. Mater Trans 2010;51:699–706. [26] Aly MS, Okuda H, Ochiai S, Fukazawa Y, Morishita K, Kato K, et al. Determination of the stress acting upon the individual cells of open-cell stainless steel foams. MetFoam 2007. In: Proc. 5th int. conf. on porous metals and metallic foams, Canada. Pennsylvania: DEStech Publications; 2008. p. 157–60. [27] Reilly DT, Burstein AH. The mechanical properties of cortical bone. J Bone Joint Surg 1974;56-A:1001–22. [28] Alvarez K, Nakajima H. Metallic scaffolds for bone regeneration. Materials 2009;2:790–832 [review]. [29] Butscher A, Bohner M, Hofmann, Gauckler L, Muller R. Structural and material approaches to bone tissue engineering in powder-based three-dimensional printing. Acta Biomater 2011;7:907–20. [30] Gibson L J, Ashby MA. Cellular solids: structures and properties. 2nd ed. Cambridge University Press; 1997. [31] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and osteogenesis. Biomaterials 2005;26:5474–91. [32] Ryan G, Pandit A, Apatsidis DP. Fabrication methods of porous metals or use in orthopaedic applications. Biomaterials 2006;27:2651–70. [33] Frosch KH, Barvencik F, Viereck V, Lohmann CH, Dresing K, Breme J, et al. Growth behaviour, matrix production, and gene expression of human osteoblasts in defined cylindrical titanium channels. J Biomed Mater Res 2004;68A:325–34. [34] Fujibayashi S, Takemoto M, Neo M, Matsushita T, Kokubo T, Doi K, et al. A novel synthetic material for spinal fusion: a prospective clinical trial of porous bioactive titanium metal for lumbar interbody fusion. Eur Spine J 2011;20:1486–95.
Contents lists available at SciVerse ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Novel multilayer Ti foam with cortical bone strength and cytocompatibility K. Kato a,⇑, S. Ochiai b, A. Yamamoto c,d, Y. Daigo e, K. Honma e, S. Matano b, K. Omori c a
Mitsubishi Material Corp., 1-297 Kitabukuro-cho, Omiya-ku, Saitama 330-8508, Japan Department of Materials Science and Engineering, Graduate School of Engineering, Kyoto University, Yoshida, Sakyo-ku, Kyoto 606-8501, Japan c Biomaterials Center, National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan d International Center for Materials Nanoarchitectonics (WPI-MANA), National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan e Mitsubishi Materials Corp., Advanced & Tools Company, High Performance Alloy Products Divison, 476 Shimoishido-shimo, Kitamoto, Saitama 364-0023, Japan b
a r t i c l e
i n f o
Article history: Received 24 August 2012 Received in revised form 31 October 2012 Accepted 13 November 2012 Available online 29 November 2012 Keywords: Porous Ti Ti foam Cortical bone Stress shielding Cytocompatibility
a b s t r a c t The major functions required for load-bearing orthopaedic implants are load-bearing and mechanical or biological fixation with the surrounding bone. Porous materials with appropriate mechanical properties and adequate pore structure for fixation are promising candidates for load-bearing implant material. In previous work, the authors developed a novel titanium (Ti) foam sheet 1–2 mm thick by an original slurry foaming method. In the present work, novel Ti foam is developed with mechanical properties compatible with cortical bone and biological fixation capabilities by layer-by-layer stacking of different foam sheets with volumetric porosities of 80% and 17%. The resulting multilayer Ti foam exhibited a Young’s modulus of 11–12 GPa and yield strength of 150–240 MPa in compression tests. In vitro cell culture on the sample revealed good cell penetration in the higher-porosity foam (80% volumetric porosity), which reached 1.2 mm for 21 days of incubation. Cell penetration into the high-porosity layers of a multilayer sample was good and not influenced by the lower-porosity layers. Calcification was also observed in the highporosity foam, suggesting that this Ti foam does not inhibit bone formation. Contradictory requirements for high volumetric porosity and high strength were attained by role-sharing between the foam sheets of different porosities. The unique characteristics of the present multilayer Ti foam make them attractive for application in the field of orthopaedics. Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction As the average age of the population is increasing in the developed world, the demand for load-bearing metals for orthopaedic implants has grown drastically [1]. The major functions required of orthopaedic implants are load bearing and mechanical or biological fixation with the surrounding bone. Ti and its alloys have been used in orthopaedic implants for load-bearing applications because of their biocompatibility, good corrosion resistance and excellent mechanical properties. However, the major problem of Ti and Ti alloy implants is a mismatch of Young’s modulus between bone (10– 30 GPa) and Ti (110 GPa), which leads to stress-shielding and results in bone resorption and implant loosening [2]. One possible solution to this problem is the development of low-modulus Ti alloys, which, over the past 20 years, has resulted in new alloys such as TNTZ (Ti–29Nb–13Ta–4.6Zr) [3,4]. Unfortunately, the Young’s modulus of human bone is too low to be reached by an alloydesigning approach. An alternative method of overcoming the mismatch of Young’s modulus is the use of porous metal coating, ⇑ Corresponding author. Tel.: +81 48 641 5111; fax: +81 48 641 6313. E-mail address: [email protected] (K. Kato).
which can reduce mismatch at the interface and achieve stable long-term biological fixation through bone ingrowths. Many types of porous metals have been developed and used to coat the surface of orthopaedic implants [5–10]. Typical fabrication methods for such porous metals are powder metallurgy [11–13], chemical vapour infiltration [14,15], space-holding [16,17], plasma spraying [18] and rapid prototyping [19–21]. Although open-pore structures have the advantages of a low Young’s modulus and bone ingrowth, leading to better fixation with bone, they have the disadvantage of insufficient mechanical strength compared with that of bulk structural materials [22]. As a result, the majority of these porous metals are applied as coatings on fully dense substrates [9]. Imwinkelried [16] reported the first application of a Ti foam device for the human lumbar spine. However, the yield strengths of this Ti foam, which had a porosity of 63–50%, were 68–140 MPa under compression. This low mechanical property limited its application to the load-bearing devices for the human lumbar spine. Therefore, porous metal fabrication technology ensuring adequate strength and tailorable open-pore structure is assumed to be very important for orthopaedic implants. To achieve these contradictory requirements of adequate strength and high porosity with adequately open-pore structure, the present authors developed novel
1742-7061/$ - see front matter Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actbio.2012.11.018
5803
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
Ti foam by stacking foam sheets of different porosities layer by layer to create multilayer Ti foam. This study examines the mechanical properties of multilayer Ti foam designed to have cortical bone strength. Cell penetration capability is evaluated by in vitro experiments and compared with a control sample prepared by stacking a single type of foam sheet. In addition, cell penetration and calcification in the Ti foam sheet are examined. 2. Materials and methods 2.1. Manufacturing process and specification of Ti foam and multilayer Ti foam An original slurry foaming method, illustrated in Fig. 1a, was used to fabricate novel Ti foam. A slurry containing Ti powder (20 lm in mean diameter), binder, foaming agent and water was prepared and coated on a carrier sheet. The sheet was treated in a foaming moisture chamber, and then moved to a drying chamber, where the foaming agent in the slurry was evaporated. Foamed green sheets were heated for debinding and then sintered in a vacuum heat-treatment furnace. This Ti foam has already been used commercially in new electrochemical applications [23]. As-fabricated Ti foam sheets were 0.2–2 mm thick and possessed a prescribed porosity and pore diameter. The volumetric porosity of the foam P was estimated by
P ¼ 1 ðq =qs Þ where q⁄ is the measured apparent density of foam, and qs is the density of bulk Ti. Because the Ti foam had a sheet shape, it was possible to stack them easily, despite their different volumetric porosities. Table 1 lists the test samples prepared by stacking the Ti foam sheets. Designation letters indicate Ti foam with high (H), middle (M) or low (L) volumetric porosity. Designation numbers indicate the approximate thickness in millimetres of the corresponding Ti foam in the stacked sheet. The ‘‘multilayer’’ type consists of two different volumetric porosity sheets, and the ‘‘stacked’’ type consists of the same volumetric porosity sheets. Table 2 lists the chemical composition of Ti powder and Ti foam. A photograph and microscopic images of H2L1, multilayer Ti foam composed of high volumetric porosity sheet (H, 80%) and low volumetric porosity sheet (L, 17%), are shown in Fig. 1b–d. As specified in Table 1, H2L1 consisted of four pairs of eight sheets: high volumetric porosity sheet with a thickness of 2 mm and low volumetric porosity sheet with a thickness of 1 mm alternated four times. The sample was sintered in a vacuum heat-treatment furnace. In the case of M2L2, M denotes a middle volumetric porosity sheet of 62%, and 2 denotes a thickness of 2 mm. Similarly, H2 was composed of a single type of 2-mm-thick high volumetric porosity Ti foam sheet. H2 data were obtained during a previous work [23].
(a) Slurry ; Ti Powder, Binder, Foaming Agent Blade Ti Foam
Carrier Sheet
Mixing
Coating
(b)
Foaming
Drying
(c)
100µm
Debinding
Sintering
(d)
10µm
Fig. 1. (a) Illustration of the slurry foaming method; (b) photograph of a multilayer Ti foam sample; (c) and (d) scanning electron microscopy images of the arrow surface of (b) in (c) low and (d) high magnification.
Table 1 Specifications of multilayer and stacked Ti foam samples. Designation
H1M1 H2L1 H1L1 M2L2 H2e M2
Type
Multilayer Multilayer Multilayer Multilayer Stacked Stacked
Total porositya (%)
72 57 42 38 80 62
No. of repetitions
6 4 6 3 6 6
Ti foam Ab
Ti foam Bb
DA (lm)
PA (%)
TA (mm)
DB (lm)
PB (%)
TB (mm)
320 320 380 340 300 340
80 80 80 62 80 62
1 2 1 2 2 2
320 –c –c –c – –
63 17 17 17 – –
1 1 1 2d – –
D, average pore diameter; P, porosity; T, thickness. Designation letters indicate Ti foam with high (H), middle (M) or low (L) volumetric porosity. Designation numbers indicate the approximate thickness of the corresponding Ti foam in the stacked sheet. a Total porosity of multilayer and stacked foam was slightly lower than the initial porosity of Ti foam A and B, because of the stacking sintering process. b Properties of Ti foam A and foam B are measured in the sheet condition. c Pore diameter of the foam with low porosity was difficult to measure by optical microscopy. d Ti foam L was stacked using two sheets of L1. e Data of stacked sample H2 was obtained in previous work [23].
5804
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
Table 2 Chemical composition of Ti powder and foam (mass%).
Ti powdera Ti foam Ab Ti foam Bc
c
H
Fe
O
N
<0.01 0.10 0.13
0.034 0.008 <0.001
0.065 0.054 0.040
0.23 0.27 0.30
0.01 0.01 0.01
Mill certificate. Chemical analyses in previous work [23]. Chemical analyses in present work.
Stacked thickness Stacked thickness
Length
a b
(b)
(a)
C
a
Designation
Average pore diametera (lm)
Porosity (%)
Dimensions (mm)
Type
Assay
H180
180
87
W12, L10, T2
Penetration
H320
320
85
W12, L10, T2
H500
500
85
W12, L10, T2
H150
150
89
W11, L11, T0.5
Single sheet Single sheet Single sheet Single sheet
Penetration
Length
Table 3 Specifications of Ti foam sheet samples.
(c) Width
Penetration Calcification
Average pore diameter was measured by digital image processing.
The pore diameter of the Ti foam was controllable between 50 and 500 lm. In the present work, four types of Ti foam sheets with different pore diameters for cell penetration and calcification assays were prepared, as shown in Table 3.
Thickness TB
TA
Fig. 2. Illustration of compression and tension test samples. Arrows show the compression and tension directions. (a) Stacked Ti foam sample; (b) multilayer Ti foam sample; (c) single Ti foam sheet. TA and TB represent the thickness of Ti foam sheets A and B in the multilayer foam.
2.2. Mechanical test samples and test procedures Mechanical test samples are listed in Table 1. Three samples of each type were tested. The compression test samples were 12 mm wide, 10 mm long and 11–12 mm in stacked thickness, as shown in Fig. 2a. The number of repetitions of foam sheet pairs was chosen to give nearly the same stacking thickness for all samples of 12 mm. The single sheet compression test sample L was 5 mm wide, 4.5 mm long and 1 mm thick, as illustrated in Fig. 2c. Compression tests were carried out at room temperature with a tensile and compression testing machine (TG-20kN, Minebea Corp, Nagano, Japan) according to JIS H 7902: Method for compressive test of porous metals. Testing was conducted at a constant crosshead speed of 0.5 mm min1. Strain measurements were carried out using two contact digital displacement sensors (AT-010, Keyence Corp, Osaka, Japan), with which the displacement of the sandwich board was accurately determined. The compression and tensile test directions were parallel to the stacking interface. The tensile test samples were 5 mm wide, 50 mm long and 11–12 mm in stacked thickness, in the directions shown in Fig. 2b. Both ends of the tensile samples were glued to aluminium tabs with epoxy-glue [23]. Tensile testing was carried out at room temperature with a tensile machine (AUTOGRAPH-AG 50kNG, Shimadzu Corp., Kyoto, Japan) at a constant crosshead speed of 8.3 106 m s1. Sample strain was measured with a non-contact extensometer [24–26]. Two paper markers were pasted onto the samples at a spacing of 20 mm, and their displacement was recorded with a laser CCD camera (DVW-200 Shimadzu Corp.). 2.3. Cell penetration assay of Ti foam sheet, stacked and multilayer Ti foam First, the effect of pore diameter on cell penetration was evaluated using the foam sheet samples (see Table 3) and, second, cell
penetration into stacked foam H2 and multilayer foam H2L1 was evaluated (see Table 1 and Fig. 3a). The stacked and multilayer samples were 2 mm wide, 10 mm long and 11–12 mm in stacked thickness. Cell penetration assays were performed in the manner shown schematically in Fig. 3b. In most cytocompatibility tests for porous materials, cell suspension is directly applied onto the porous samples. As a result, some of the cells stay at the very top surface and some drop through the pores, and thus inoculation itself causes some difference in cell penetration distance at the very first stage of cell culture. To avoid such error in this study, human osteosarcoma Saos-2 cells were seeded on a sterile silicone sheet (15 mm square and 0.1 mm thick) at a concentration of 1.85 105 cells per 6 ml of Dulbecco’s modified minimum essential medium supplemented with 10 vol.% foetal bovine serum (D-MEM + FBS), cultured for 4 days to be confluent, and the resulting cell-covered sheet was then attached to the autoclaved specimen surface, as shown in Fig. 3b. A glass ring 9 mm in diameter was placed over the silicone sheet to prevent movement of the sheet. The cells were cultured for 7, 14 and 21 days, while the culture medium was replaced with fresh medium every other day. Then, the cells were fixed with a mixture of 10 vol.% formalin and 10 vol.% methanol. The sample was sliced longitudinally with a diamond cutter, as shown in Fig. 3b, and the cells were stained with sulforhodamine 101 acid chloride. Reproducibility was confirmed by repeating the same experiment. Cells in the sample were observed from the side (longitudinal plane) with a fluorescent microscope (MVX-10, Olympus Co, Ltd., Tokyo, Japan). The farthest distance from the specimen surface to cells in a view field of 3.52 2.68 mm was measured and averaged for more than four view fields per sample. The statistical significance of data obtained for three different single sheets was analysed by the Tukey (or Tukey–Kramer) test, while those for
5805
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
(a) 1mm
Subsequently, the culture medium was replaced with calcification medium; D-MEM + FBS containing 0.5 mM b-glycerophosphate and 50 lg ml1 L-ascorbic acid. The cells were cultured for 7, 14, 21 and 28 days, while the medium was exchanged every other day. Calcein was then added to give a concentration of 1 lg ml1, and the microplate was kept in the CO2 incubator for a further 4 h. Then, the cells were fixed and stained in the manner described above. The cells in the sample were observed by a confocal laser scanning microscope (LSM-510, Carl Zeiss Japan, Tokyo, Japan) for more than three views for one sample. The reproducibility was confirmed by repeating the same experiment. 3. Results 3.1. Mechanical properties of the multilayer Ti foam
(b)
cell
4d of incubation
metal foam 7, 14, or 21d of incubation
Fixation Cutting Fluorescent staining
Observation by fluorescent microscopy
The compression and tensile test results are listed in Table 4. In compression tests, the Young’s modulus and yield strength of multilayer samples with L1 or L2 were 11 GPa and 150–250 MPa, respectively. It is notable that these results for the present multilayer Ti foam with L1 or L2 are comparable with those for cortical bone, the Young’s modulus and strength of which under compression are 10–30 GPa and 150–220 MPa [27–29], respectively. In tensile tests, the Young’s modulus and strength of multilayer samples with L1 or L2 were 17–22 GPa and 100–110 MPa, respectively. These results are also comparable with those of cortical bone, the Young’s modulus and strength of which are 10–25 GPa and 90–170 MPa, respectively [27–29]. Tensile Young’s moduli were higher than the compression moduli, because of uniform stress distribution in the foam. In tensile tests, the grip section of the sample was glued to avoid stress concentration and, furthermore, stress–strain curves were measured in the gauge section
Cell Penetration Distance (mm)
silicone sheet
Fig. 3. (a) Photographs of multilayer Ti foam sample H2L1 for cell penetration assay and (b) schematic explanation of cell penetration assay procedure.
stacked and multilayer foam were analysed by an unpaired t-test (or Welch’s modified t-test). 2.4. Calcein assay of Ti foam sheet A foam sheet sample 11 mm square and 0.5 mm thick was set into a silicone O-ring to prevent its direct contact with the bottom of the culture vessel surface to avoid penetration of cells growing on the culture vessel surface into the foam sheet. The foam sheet in the silicone O-ring was autoclaved and then placed on the well of a ‘‘non-treated’’ 12-well microplate. Saos-2 was inoculated onto the sample at a concentration of 1 106 cells in 0.1 ml DMEM + FBS. After 20 min, a 2 ml portion of D-MEM + FBS was added, and the cells were then cultured for 3 days to be confluent.
1.5
H180 H320 H500
1.0
0.5
*
mean s.d.
0 0
5
10
15
20
25
Incubation Period (d) Fig. 4. Cell penetration into Ti foam sheets with different open-pore sizes. Subscripts in graph indicate mean pore diameter (see Table 2). ⁄Significant difference is only observed between H320 and H500 after 7 days’ incubation (p < 0.05).
Table 4 Compresion and tensile test results for multilayer and stacked Ti foam samples. Designation
Porosity (%)
Compression yield strength (MPa)
Compression Young’s modulus (GPa)
Tensile strength (MPa)
Tensile Young’s modulus (GPa)
Nominal tensile fracture strain of tensile test (%)
H1M1 H2L1 H1L1 M2L2 H2a M2 Lb
72 57 42 38 80 62 17
34.3 ± 0.6 158.3 ± 14.6 240.0 ± 10.0 251.7 ± 2.9 19.4 ± 4.1 47.8 ± 0.2 431.0 ± 28.2
2.8 ± 0.2 11.1 ± 2.3 11.5 ± 1.4 11.1 ± 0.5 3.2 ± 0.6 6.4 ± 0.2 17.1 ± 1.1
25.5 ± 1.8 110.1 ± 19.3 103.7 ± 4.9 113.8 ± 6.5 20.2 ± 2.1 – –
5.8 ± 1.2 21.7 ± 6.7 17.6 ± 1.2 19.1 ± 2.7 3.8 ± 1.2 – –
0.98 ± 0.31 0.72 ± 0.15 0.68 ± 0.17 0.58 ± 0.05 1.35 ± 0.86 – –
Test results are mean ± standard deviation. Designation letters indicate Ti foam with high (H), middle (M) or low (L) volumetric porosity. Designation numbers indicate the approximate thickness of the corresponding Ti foam in the stacked sheet. a Obtained in previous work [23]. b Single Ti foam sheet. Compression was loaded in the direction of the sheet thickness, assuming isotropy of the foam sheet L.
5806
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
Stacked foam (H2)
Multilayer Ti foam (H2L1) cell contacting surface
7d
0.5mm
0.5mm
0.5mm
0.5mm
0.5mm
0.5mm
14d
21d
0.5mm
0.5mm
Cell Penetration Distance (mm)
Fig. 5. Fluorescence microscopy images of cells penetrating samples H2 and H2L1. Arrows indicate furthest cell penetration distance in the imaged area. The bottom images show longitudinal cross-sections of both samples.
1.5 H2 H2L1 1.0
0.5 mean s.d. 0
0
5
10
15
20
25
Incubation Period (d) Fig. 6. Cell penetration into samples H2 and H2L1. No significant difference was observed between H2 and H2L1 for all incubation periods.
apart from the grapping portion. Because of this specially designed test method, tensile stress was distributed uniformly in cross section. In the compression test, the sample foam surface contact to the compression tool plate was relatively uneven, which did not give uniform stress in cross section. In contrast to the compression
yield strength, the tensile strength did not increase with decreasing porosity, as shown in Table 4. This difference may be attributed to the lack of uniform pore dispersion; in tensile testing, any irregular porosity brings about stress concentration, which causes early failure. In Fig. 8, the compression yield strength data in Tables 4 and 5 are plotted against the porosity. Two models were applied to these results to evaluate the effect of stacking low-porosity and highporosity layers on the yield strength of the multilayer foam: the Gibson–Ashby model [30] (Eq. (4.1) shown later in Section 4.1) and the rule of mixtures (Eq. (4.2) shown later in Section 4.1). The curved line calculated by Eq. (4.1) and the straight lines calculated by Eq. (4.2) are also shown in Fig. 8. 3.2. Cell penetration in the Ti foam sheet, stacked and multilayer Ti foam Fig. 4 presents the results of the cell penetration assay for the three types of Ti foam sheet with different pore diameters. In comparison with H180, H320 had larger cell penetration distances for all
5807
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
7d
alongside the cells, suggesting the importance of cell penetration into the Ti foam.
14d
4. Discussion 4.1. Mechanical properties
100µm
100µm
21d
28d
100µm
100µm
28d
Most mechanical tests of Ti foam in the literature have been by compression testing. Therefore, the compression properties of H1L1, produced by the slurry foaming method, were compared with those of samples produced by other methods, with volumetric porosities ranging from 33% to 50%, as shown in Table 5. H1L1 had a superior 0.2% yield strength compared with the other samples. The reason for this stems from the existence of the 17% porosity layer, which had a high yield strength of 431 MPa, as measured in the present work (see Table 4). For reference, the 0.2% yield strength of the stacked sample of 80% porosity sheets was 19.4 MPa, estimated in recent work [23] (also shown in Table 4). In Fig. 8, the concrete form of the equation for each model is summarized as follows, with important indications obtained by comparison of the experimental data with the calculations. According to Gibson–Ashby [30], the yield strength (rpl ) of open-pore foam under compression can be expressed as a function of the relative density (q⁄/qs), with the apparent densities of the foam q⁄ and the solid material qs.
rpl =rys ¼ C ðq =qs Þ3=2
ð4:1Þ
In Eq. (4.1), rys and C are the yield strength of the solid material and a constant of proportionality, respectively. Substituting the yield strength rys = 650 MPa [16] and the test results obtained for H2 in into Eq. (4.1), one has C = 0.33, which is in good agreement with the 0.3 used in Gibson–Ashby’s work [30]. However, one can draw straight lines between L and H2 and between L and M2. These lines can be expressed using the volumetric ratio method given by Eq. (4.2), in which the yield strength rm of multilayer foam is approximated to be given by the rule of mixtures for the yield strength values of the constituting foam sheets.
100µm
rm ¼ rA TA =ðTA þ T B Þ þ rB TB =ðTA þ T B Þ
Fig. 7. Calcification of Saos-2 on a single Ti foam sheet, H150. Green indicates densely calcified areas marked by calcein, and orange indicates cells marked mainly by Texas Red (and slightly by calcein as well).
incubation periods, reaching 1.2 mm after 21 days’ incubation. In contrast, H500 showed a cell penetration distance similar to that of H320 for 7 and 14 days’ incubation, but remained 1.0 mm for 21 days’ incubation. Results for the stacked and multilayer samples, H2 and H2L1, are shown in Figs. 5 and 6. Cell penetration distance tended to increase with longer incubation period, and no difference between H2 and H2L1 was observed. This confirmed that maximum cell penetration into the multilayer sample was obtained at the higher porosity layers and was not influenced by the lower porosity layers.
3.3. Calcification in the Ti foam sheet The results of the calcification assay for H150 are shown in Fig. 7, in which calcification of Saos-2 in the Ti foam was observed. The densely calcified area clearly increased with incubation period, and was especially apparent after 28 days’ culture. The magnified images at the bottom of Fig. 7 revealed that calcification occurred
ð4:2Þ
Here, r and r are the yield strengths, and TA and TB the thicknesses of foam A and foam B, respectively. The following conclusions are read from the experimental data and calculation results shown in Fig. 8. (1) The curved line of Eq. (4.1) approximately describes the experimental data obtained for samples with relatively low yield strength; the single or stacked samples and multilayer samples composed of relatively high-porosity foam. In contrast, the volumetric ratio method ((4.2)) approximately describes the data obtained for the multilayer samples. This finding suggests that the mechanism controlling the yield strength of multilayer samples with closed pore foam was different from that of those with uni A
B
Table 5 Compression test results for Ti foam produced by slurry foaming, rapid prototyping, space holder, and plasma-spray methods. Methods
Porosity (%)
Pore diameter (lm)
Yield strength (MPa)
Young’s modulus (GPa)
Ref.
Slurry foaming, multilayer Ti foam Rapid prototyping Space holder Plasma spray
42
380
230–250
10–13
Present work
33–40
700
100–210
13–20
[20]
50 40
100–500 300–500
130–140 85
16–18 4–6
[16] [18]
5808
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
500
450 0.2% Yield Strength, MPa
Present work with porosity 17% layers Present work without porosity 17% layers Porosity 17% sheet Previous work [23] Ref erence data
L sheet
400
350 300 250
H1L1
M2L2
200
[20]
H2L1
150
Gibson-Ashby 100 Equation curve
[16] [18]
50
H2
M2 H1M1
0 0
10
20
30
40
50
60
70
80
90
Porosity, % Fig. 8. Compression 0.2% yield strength plotted against porosity.
form open-pore foam. (2) The straight line ((4.2)) for the foam containing 17% porosity L sheets exhibited a higher yield strength than the curved line of Eq. (4.1) for open-pore foam. This means that the multilayer Ti foam had a yield strength superior to porous materials, whose yield strength followed Eq. (4.1). (3) The 0.2% yield strength of the present multilayer Ti foam is thus controllable by changing the volumetric ratio of higher-porosity foam A to lower-porosity foam B, as shown in Table 1. 4.2. Cell penetration In vivo implantation tests into animal bone tissue are generally performed to evaluate the biological fixation ability of porous materials. Such testing takes a long period of time, is costly and has limitations on the number of implanting samples. Most importantly, differences in species between animals and humans often bring different results. Therefore, animal experiments are expected to be refined, reduced and replaced by other evaluation methods such as in vitro experiments. For bone formation, either body fluid or cells must penetrate into the foam, which is influenced by the foam structure. Using appropriate experimental conditions, it is possible to evaluate cell penetration into the foam by in vitro experiments as a screening test. However, studies on cytocompatibility of porous Ti or other metals with different pore structures have only evaluated cell proliferation. No research has been performed as yet to evaluate cell penetration into metal foam. Therefore, cell penetration was evaluated for different pore diameters. It was found that H320 had the most stable and largest cell penetration distance among the foam sheets tested. This may be because of a balance between the opportunity for cells to transfer from the silicone sheet to the pore walls and penetration efficiency for deeper penetration into the foam. A larger pore diameter can offer higher efficiency to gain penetration distance, but also less opportunity for cells to transfer to the foam. H320 was considered to have an optimal balance between penetration efficiency and transferral opportunity among the foam sheets tested in the cell penetration assay. Based on in vivo experiments, 100 lm is generally considered the minimum recommended pore size for bone ingrowth, but pores >300 lm have better bone formation, because relatively
larger pores favour direct osteogenesis because of their higher vascularization and oxygenation [31]. Pore sizes >1 mm, however, increase fibrous tissue formation [32]. The pore size of H320 agreed with the favourable range described above. Therefore, foam of this diameter was employed in the preparation of stacked and multilayer samples to support cell penetration. Cell penetration assays revealed that multilayer samples including this foam were able to maintain good cell penetration in the higher-porosity layers with no influence of the lower-porosity layers. Combining Ti foam sheets with different porosities or pore structures while maintaining relatively high cell penetration increases the range of orthopaedic load-bearing device designs. The present multilayer Ti foam, with a combination of lower- and higher-porosity foam sheets, exceeds the limitations of previous foam materials by sharing the contradictory roles of strength and cell penetration.
4.3. Cell calcification Calcification is one of the steps occurring at the early stage of bone formation in vivo, and is a good measure of bone tissue mineralization. Although in vitro calcification assay is often performed to examine cellular differentiation and maturity of osteogenic cells, it can also be used to evaluate the ability of materials to support the differentiation and mineralization of cells cultured on them. In the present study, calcification of Saos-2 in the single Ti foam sheet was confirmed, as shown in Fig. 7. Densely calcified areas were observed alongside the cells, suggesting the importance of cell penetration into Ti foam for bone formation. It has been reported that, for a Ti specimen with drilled channels of different diameters, the amount of mineralization of human osteoblasts in the channels correlates well with the cell penetration distance [33]. These results indicate the effectiveness of cell penetration assay as an in vitro evaluation method to estimate bone formation in porous materials. Thus, based on the observed good cell penetration and evidence of calcification, this Ti foam was expected to be suitable for implant applications requiring fixation with bone.
K. Kato et al. / Acta Biomaterialia 9 (2013) 5802–5809
4.4. Application Spacers, such as interbody fusion cages for the human lumbar spine, are one typical orthopaedic load-bearing device. The current major materials used for these cages are solid Ti and Ti alloys, because of their high strength. However, failed fusion and cage malposition or migration have been identified in such devices [9]. Recently, Fujibayashi et al. [34] reported a novel porous Ti sintered with a solid outer frame for clinical use, which achieved successful spine fusion within 6 months. They explained the importance of spacer materials with high strength and high bone-bonding ability for spinal fusion without autologous iliac crest bone grafting [34]. The present multilayer Ti foam has adequate Young’s modulus, high yield strength in compression because it has low-porosity layers, and good cell penetration because it also has high-porosity layers. Furthermore, its high yield strength compared with other porous materials can be controlled by changing the thickness ratio of the high-porosity layer to low-porosity layer. These characteristics of multilayer Ti foam make it promising for load-bearing implant materials in the field of orthopaedic spinal devices. 5. Conclusions A novel multilayer Ti foam with mechanical properties compatible for cortical bone and biological fixation capability was developed. The Young’s modulus and yield strength of the foam were controllable to the range of those exhibited by cortical bone. Good cell penetration and calcification was also observed during in vitro screening tests. Cell penetration assay of a single Ti foam sheet showed that Ti foam with a 320 lm pore diameter and 85% volumetric porosity had optimum cell penetration distance among the three foam sheets tested in this study. A cell penetration assay for a multilayer sample containing this foam sheet revealed good cell penetration in the higher porosity layers without any influence of the lower porosity layers. Calcification of Saos-2 was also confirmed in the single Ti foam sheet. The multilayer Ti foam solves the contradictory requirements of adequate strength and high porosity with adequate open-pore structure. The unique characteristics of multilayer Ti foam make it an attractive load-bearing implant material, especially in the field of spinal orthopaedics. Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 1, 3, 5 and 7, are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/ j.actbio.2012.11.018. References [1] Nasab MB, Hassan MR, Sahari BB. Metallic biomaterials of knee and hip – a review. Trend Biomater Artif Organs 2010;24:69–82. [2] Krishna BV, Bose S, Bandyopadhyay A. Low stiffness porous Ti structures for load-bearing implants. Acta Biomater 2007;3:997–1006. [3] Akahori T, Niinomi M, Koyanagi Y, Kasuga T, Toda H, Fukui H, et al. Mechanical properties of biocompatible beta-type titanium alloy coated with calcium phosphate invert glass-ceramic layer. Mater Trans 2005;46:1564–9. [4] Tutumi H, Niinomi M, Akahori T, Nakai M, Takeuchi T, Katsura S. Mechanical properties of a beta-type titanium alloy cast using a calcia mold for biomedical applications. Mater Trans 2010;51:136–42. [5] Hahn H, Palich W. Preliminary evaluation of porous metal surfaced titanium for orthopedic implants. J Biomed Mater 1970;4:571–7. [6] Bobyn JD, Pilliar RM, Cameron HU, Weatherly GC. The optimum pore size for the fixation of porous- surfaced metal implants by the ingrowth of bone. Clin Orthop Relat Res 1980;150:263–70.
5809
[7] Levine B. A new era in porous metal: applications in orthopedics. Adv Eng Mater 2008;10:788–92. [8] Wazen RM, Lefebvre LP, Baril E, Nanci A. Initial evaluation of bone ingrowth into a novel porous titanium coating. J Biomed Mater Res B 2010;94B:64–71. [9] Singh R, Lee PD, Dashwood RJ, Lindley TC. Titanium foams for biomedical applications: a review. Mater Technol 2010;25:127–36. [10] Mour M, Das D, Winkler T, Hoenig E, Mielke G, Morlock MM, et al. Advances in porous biomaterials for dental and orthopedics application – review. Materials 2010;3:2947–74. [11] Oh IH, Nomura N, Masahashi N, Hanada S. Mechanical properties of porous titanium compacts prepared by powder sintering. Scripta Mater 2003;49:1197–202. [12] Chen J, Paetzell E, Zhou J, Lyons L, Soboyejo W. Osteoblast-like cell ingrowth, adhesion and proliferation on porous Ti–6Al–4V with particulate and fiber scaffolds. Mater Sci Eng 2010;C30:647–56. [13] Torres Y, Pavon JJ, Nieto I, Rodriguez JA. Conventional powder metallurgy process and characterization of porous titanium for biomedical applications. Metall Trans B 2011;42B:891–900. [14] Zardiackas LD, Parsell DE, Dillon LD, Mitchell DW, Nunnery LA, Poggie R. Structure, metallurgy, and mechanical properties of a porous tantalum foam. J Biomed Mater Res Appl Biomater 2001;58:180–7. [15] Shimko DA, Shimko VF, Sander ED, Dickson KF, Nauman EA. Effect of porosity on the fluid flow characteristics and mechanical properties of tantalum scaffolds. J Biomed Mater Res 2005;73:315–24. [16] Imwinkelried T. Mechanical properties of open-pore titanium pore. J Biomed Mater Res 2007;81A:964–70. [17] Hong TF, Guo ZX, Yang R. Fabrication of porous titanium scaffold materials by a fugitive filler method. J Mater Sci Mater Med 2008;19:3489–95. [18] Takemoto M, Fujibayashi S, Neo M, Suzuki J, Kokubo T, Nakamura T. Mechanical properties and osteoconductivity of porous bioactive titanium. Biomaterials 2005;26:6014–23. [19] L-Heredia MA, Sohier J, Gaillard C, Quillard S, Dorget M, Layrolle P. Rapid prototyped porous titanium coated with calcium phosphate as a scaffold for bone tissue engineering. Biomaterials 2008;29:2608–15. [20] Ryan GE, Pandit AS, Apatsidis DS. Porous titanium scaffolds fabricated using a rapid prototyping and powder metallurgy technique. Biomaterials 2008;29:3625–35. [21] Shishkovskii IV, Yadroitsev IA, Smurov. Selective laser sintering/melting of nitinol–hydroxyapatite composite for medical application. Powder Metall Met Ceram 2011;50:275–83. [22] Balla VK, Bodhak Z, Bose S, Bandyopadhaya A. Porous tantalum structures for bone implants: fabrication, mechanical and vitro biological properties. Acta Biomater 2010;6:3349–59. [23] Kato K, Yamamoto A, Ochiai S, Daigo D, Isobe T, Matano S, et al. Cell proliferation, corrosion resistance and mechanical properties of novel titanium foam with sheet shape. Mater Trans 2012;53:724–32. [24] Ochiai S, Nakano S, Fukazawa Y, Aly MS, Okuda H, Kato K, et al. Change of Young’s modulus with increasing applied tensile strain in open cell nickel and copper foams. Mater Trans 2010;51:925–32. [25] Ochiai S, Nakano S, Fukazawa Y, Aly MS, Okuda H, Kato K, et al. Tensile deformation and failure behavior of open cell nickel and copper foams. Mater Trans 2010;51:699–706. [26] Aly MS, Okuda H, Ochiai S, Fukazawa Y, Morishita K, Kato K, et al. Determination of the stress acting upon the individual cells of open-cell stainless steel foams. MetFoam 2007. In: Proc. 5th int. conf. on porous metals and metallic foams, Canada. Pennsylvania: DEStech Publications; 2008. p. 157–60. [27] Reilly DT, Burstein AH. The mechanical properties of cortical bone. J Bone Joint Surg 1974;56-A:1001–22. [28] Alvarez K, Nakajima H. Metallic scaffolds for bone regeneration. Materials 2009;2:790–832 [review]. [29] Butscher A, Bohner M, Hofmann, Gauckler L, Muller R. Structural and material approaches to bone tissue engineering in powder-based three-dimensional printing. Acta Biomater 2011;7:907–20. [30] Gibson L J, Ashby MA. Cellular solids: structures and properties. 2nd ed. Cambridge University Press; 1997. [31] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and osteogenesis. Biomaterials 2005;26:5474–91. [32] Ryan G, Pandit A, Apatsidis DP. Fabrication methods of porous metals or use in orthopaedic applications. Biomaterials 2006;27:2651–70. [33] Frosch KH, Barvencik F, Viereck V, Lohmann CH, Dresing K, Breme J, et al. Growth behaviour, matrix production, and gene expression of human osteoblasts in defined cylindrical titanium channels. J Biomed Mater Res 2004;68A:325–34. [34] Fujibayashi S, Takemoto M, Neo M, Matsushita T, Kokubo T, Doi K, et al. A novel synthetic material for spinal fusion: a prospective clinical trial of porous bioactive titanium metal for lumbar interbody fusion. Eur Spine J 2011;20:1486–95.