J Mater Sci: Mater Med DOI 10.1007/s10856-014-5140-5
Porous silicon confers bioactivity to polycaprolactone composites in vitro J. R. Henstock • U. R. Ruktanonchai L. T. Canham • S. I. Anderson
•
Received: 19 September 2013 / Accepted: 2 January 2014 Ó Springer Science+Business Media New York 2014
Abstract Silicon is an essential element for healthy bone development and supplementation with its bioavailable form (silicic acid) leads to enhancement of osteogenesis both in vivo and in vitro. Porous silicon (pSi) is a novel material with emerging applications in opto-electronics and drug delivery which dissolves to yield silicic acid as the sole degradation product, allowing the specific importance of soluble silicates for biomaterials to be investigated in isolation without the elution of other ionic species. Using polycaprolactone as a bioresorbable carrier for porous silicon microparticles, we found that composites containing pSi yielded more than twice the amount of bioavailable silicic acid than composites containing the same mass of 45S5 Bioglass. When incubated in a simulated body fluid, the addition of pSi to polycaprolactone significantly increased the deposition of calcium phosphate. Interestingly, the apatites formed had a Ca:P ratio directly proportional to the silicic acid concentration, indicating that silicon-substituted hydroxyapatites were being spontaneously formed as a first J. R. Henstock (&) Institute for Science and Technology in Medicine, Keele University, Stoke-on-Trent ST4 7QB, UK e-mail: [email protected] U. R. Ruktanonchai National Nanotechnology Center (NANOTEC), National Science and Technology Development Agency (NSTDA), Thailand Science Park, Pathumthani 12120, Thailand L. T. Canham pSiMedica Ltd., Malvern Hills Science Park, Geraldine Road, Malvern, Worcestershire WR14 3SZ, UK S. I. Anderson Royal Derby Hospital Centre, School of Graduate Entry Medicine and Health, University of Nottingham, Uttoxeter Road, Derby DE22 3DT, UK
order reaction. Primary human osteoblasts cultured on the surface of the composite exhibited peak alkaline phosphatase activity at day 14, with a proportional relationship between pSi content and both osteoblast proliferation and collagen production over 4 weeks. Culturing the composite with J744A.1 murine macrophages demonstrated that porous silicon does not elicit an immune response and may even inhibit it. Porous silicon may therefore be an important next generation biomaterial with unique properties for applications in orthopaedic tissue engineering.
1 Introduction Silicon is an essential element in human nutrition, with the symptoms of a silicon-deficient diet including abnormal bone development [1–4]. Similarly, animals and bone forming osteoblast-like cells in vitro show an increase in osteogenesis when supplemented orthosilicic acid [5–7] the only bioavailable source of silicon for higher animals [8]. Research has shown that supplementing dietary silicon content in humans is associated with improved bone mineral density (BMD), with supplementation of 40 mg/day increasing cortical hip BMD by up to 10 % [9]. Silicic acids are released from Bioglass and confer bioactivity by polycondensing in vivo to form a hydrated silica gel layer which adsorbs growth factors, supports cell growth and acts a nucleation bed for bone mineral, facilitating a strong chemical bond to living bone [10]. Porous silicon (pSi) exhibits several of the same characteristics as bioactive glass by forming a surface layer of hydroxyapatite in simulated body fluid resulting from the release of silicic acids in aqueous solution and is therefore of interest as a potential bioactive material [11]. pSi is made by the electrochemical etching of polycrystalline wafers, creating
123
J Mater Sci: Mater Med
nanoscale silicon hydride-lined pores through the material and resulting in an aqueously soluble material with a large surface area to volume ratio (500 m2/cm3) and hence high reactivity [11]. Porous silicon has many potential uses in the fields of opto-electronics and an emerging number of biological and clinical applications. The long, narrow pores can be filled with pharmaceuticals, creating a dissolvable drug delivery material with release kinetics that are easily controllable by adjusting the pore morphology and drug loading density [12, 13]. pSi is also being studied for applications in radioisotope intratumoral brachytherapy for treatment of cancers such as pancreatic and hepatocellular carcinoma [14]. It has been shown that injected pSi particles are fully degraded into silicic acid and cleared by excretion via the kidneys [15]. Drug delivery using dispersible pSi particles has been explored by several research groups with the release kinetics of various molecules including cis-platin [16], dexamethasone [12, 17], ibuprofen [18], and doxorubicin [19] having been reported. A comprehensive review of pSi for drug delivery applications is given in Salonen et al. [20]. Porosified silicon wafers have been shown to have low cytotoxicity, supporting the growth in vitro of several cell types including osteoblasts [21], neurones [22], and hepatocytes [23]. However, like bioactive glasses, the mechanical properties of high porosity silicon are poor, and so for these experiments we developed a composite material to explore the bioactive properties of pSi within a strong but resorbable polycaprolactone matrix, a polymer which has existing uses in tissue engineering [24], drug delivery [25], and is FDA approved for human in vivo use [26]. In this investigation, we have evaluated a range of polycaprolactone composites containing 0.4–12 % pSi by weight and characterised them by their formation of silicic acids, silica gel and hydroxyapatite in simulated body fluid. The aims of this research are initially to assess pSi as a bioactive material with unique properties suitable for inclusion in polymer-based composites for orthopaedic applications.
2 Materials and methods 2.1 Porous silicon Porous silicon for this investigation was produced by the anodisation of an electronic-grade silicon wafer (200 mm diameter, 725 lm thick) using a platinum cathode and immersed in a hydrogen fluoride electrolyte. A direct current is passed through the cell which corrodes the anode wafer, producing an even, porous layer which was dried, milled and filtered to produce 5–35 lm pSi particles with
123
an average diameter of 11 lm (Fig. 1a). A comprehensive methodology for silicon porosification using this method is given in Halimaoui [27]. 2.2 Composite formation A composite of polycaprolactone and pSi microparticles was made by dissolving the polymer in a small volume of acetone at 40 °C and mixing in pSi particles in weight percentages of 0.4–12 %. The acetone was evaporated at room temperature in a flow hood to create a solid composite which was cut into sections and placed in a 5 ml polypropylene syringe body. After heating to 70 °C to melt together the composite the cylinder was cooled until solid, removed from the syringe body and cut into 10 mm diameter 250 mg discs containing 1 (0.4 % w/w) to 30 mg (12 % w/w) silicon which were used for subsequent testing and compared to polycaprolactone discs made in the same way without pSi. To visualise the particle distribution, 1 lg/ml fluorecein solution was adsorbed into to the pSi particles prior to inclusion by brief immersion followed by filtration. 2.3 Mechanical testing The tensile strength of the composite was tested using a cylinder of the material cast in a 1 ml polypropylene syringe body. Polycaprolactone casts were tested against 4 and 8 % pSi–PCL composites. The samples were fitted into a 50 KN load cell in an Instron 5569 Tensile Testing System to test a 20 mm region of the sample. Samples were fitted into grips which were moved apart at 2.0 mm/min until fail. Maximum load and tensile stress at maximum load were measured. Screw-shaped composites were produced by melting the material inside a stainless steel threaded mould (internal diameter 12 mm, length 14 mm) on a hotplate at 80 °C and then cooling the cast inside the mould until solid. The steel mould was held in a recessed block over a central hole and a 6 mm diameter steel dowel clamped in the upper grip of the Instron testing system. The dowel was lowered to touch the polymer/composite surface and then pushed out through the mould at 2 mm/min. The force (N) required to push out the casting was measured for PCL, 4 and 8 % composites and the failure mode determined by SEM of sections through the pushed-out material. 2.4 Bioactivity assays The primary test for bioactivity involved immersing the composite in simulated body fluid (SBF), prepared as described by Kokubo et al. [28]. Composite discs were incubated for periods of 28–90 days and compared to polycaprolactone as a control. pSi was compared to
J Mater Sci: Mater Med
A
B
C
D
Fig. 1 a The 70 % porous silicon microparticles used in this investigation, stained with fluorescein and imaged using a confocal microscope showing the even distribution of 11 lm particles within the polycaprolactone composite. b Typical stress–strain curve of composite (8 % pSi) showing the locations on the curve of the proportionality limit beyond which stress/strain is non-linear and
below which Young’s modulus is calculated, the ultimate (or yield) stress, and the beginning and mode of failure. c Elastic modulus was calculated from the gradient of the stress–strain curve below the elastic limit. d The force required to push out the casting from the mold using a metal dowel. Error bars are standard error of the mean, n = 4. Scale bar 100 lm
bioactive glass by forming composites made from polycaprolactone with 8 % (w/w) either 45S5 Bioglass particles or pSi and incubating in water for 42 days.
weight before incubation and after 90 days, compared to polycaprolactone-only controls. Confirmation of silicia gel and apatite formation was by scanning electron microscopy and X-ray microanalysis (EDX) in samples imaged using a Phillips XL30 FEG-ESEM. After incubation for 28 days in SBF, the discs were rinsed, dried and imaged at 10 kV in auxiliary mode with nitrogen as the column gas allowing energy dispersive X-ray microanalysis to be performed on uncoated samples. The elemental composition of surface layers was mapped using the integrated elemental mapping software. Additional SEM (e.g. on castings for mechanical testing) was performed on samples which were gold coated and imaged using a Jeol JSM840 scanning electron microscope.
2.5 Silicon degradation Analysis of silicic acid in the above solutions (resulting from the degradation of pSi) was by a molybdenum (molybdate) blue assay, in which ammonium molybdate in solution reacts with silicic acid in the sample to form bsilicomolybdic acid (yellow) which is subsequently reduced to form a molybdate blue complex with a peak absorbance at 810 nm. Experimental values for soluble silicon species (ortho- and polysilicic acids) were calibrated against known concentrations of sodium metasilicate.
2.7 Macrophage cell culture 2.6 Silica gel and apatite formation Polycondensation of silica on the material was evaluated by incubating the composites in dH2O and comparing dry
The murine BALB/c monocyte macrophage cell line J774A.1 was used to determine the effect of the composite on macrophage morphology and activation. Composite
123
J Mater Sci: Mater Med
samples of PCL containing 0.5, 1.5, 3 or 4.5 % pSi (equating to 1, 3, 6 or 9 mg/200 mg composite disc) were placed in the wells of a 24 well plate (Nunc). Thermanox tissue culture plastic coverslips and PCL polymer samples were used as controls. Cells were seeded at a density of 2 9 105 cells/ml in Dulbecco’s Modified Eagles Medium (DMEM) containing 10 % FCS, 5 % sodium bicarbonate, 1 % glutamine and 1 % penicillin/streptomycin and incubated at 37 °C in a humidified atmosphere containing 5 % CO2 for 48 h. After 48 h, samples were washed in warmed sterile PBS and fixed in 4 % paraformaldehyde for 5 min. Samples were permeabilised in a solution containing triton X100 for 10 min at 0C and the f-actin was stained with FITC-conjugated phalloidin (Sigma) for 20 min. The cell nuclei were then stained with PI for 30 s. Samples were cover-slipped and viewed in a Leica TCS-4D confocal laser scanning microscope (CLSM) using the 488 nm laser line to excite the green fluorescence of FITC-conjugated phalloidin and the 568 nm laser line to excite the red fluorescence of PI. Optical sections were taken at 0.3 mm intervals and displayed as a maximum intensity projection using Scanware software.
in DMEM supplemented with 10 % FCS, 2 mM L-glutamine, 1 % non-essential amino acids, 0.2 mM HEPES, 150 lg/ml ascorbic acid, 2 % penicillin/streptomycin and with 10 mM sodium b-glycerophosphate and 10-7 M dexamethasone (all Sigma, UK). Osteoblasts were seeded at a density of 80,000 cells/ml directly onto the polycaprolactone or composite disc surface or thermanox coverslip (Nunc) in 24-well plates (Nunc). Media was replaced every 48 h and the osteoblasts incubated at 37 °C in 5 % CO2. Cell lysates (n = 6) were obtained at intervals of 7, 14, 21 and 28 days by freezing the monolayer cells in dH2O. DNA was quantified by incubating cell lysate samples with 1 % bisBenzimide H (Hoechst 33258) dye (Sigma, UK). Samples were measured and the fluorescence was read at excitation 355 nm and emission 460 nm using a Fluoroskan Ascent reader and compared to a standard curve of calf thymus DNA (Sigma, UK). Alkaline phosphatase (ALP) was quantified using p-nitrophenyl phosphate (Sigma, UK) which forms a complex with ALP present in the cell lysate with a peak absorbance read at 405 nm. Collagen was quantified in cell lysates using the sirius red-based Sircol collagen assay kit (Biocolor, UK) at 540 nm and compared to a standard curve of type-I rat tail collagen.
2.8 Macrophage activation 2.10 Statistics Macrophage activation was quantified using the Quantikine M Mouse IL-1b immunoassay (ELISA) (R&D Systems) to determine the levels of IL-1b produced and secreted into the culture medium by macrophages growing on the different surfaces. Samples were prepared and cells seeded as above. Macrophages grown on tissue culture plastic (TCP) were used as the negative control and macrophages grown on TCP containing 10 or 100 mg/ml lipopolysaccharide (LPS), a known macrophage activator, were used as the positive control. After 48 h the medium was removed from the wells and the assayed with the absorbance measured at 450 nm for test samples and known concentrations of IL1b. Quantities of IL-1b were determined by interpolation from the standard curve. The production of hydrogen peroxide by J774A.1 macrophages seeded on composite surfaces was measured using dichlorofluorescein (DCF). Cells were seeded at a concentration of 2 9 105 cells/ml as before. TCP and PCL were used as controls. Copper substrates were also used as positive controls. Cells were incubated in DCF (1:100 in Hank’s Balanced Salt Solution). PMA (Phorbol 12-Myristate 13-Acetate, Sigma) was used as a positive control and was added to the appropriate wells at 15 ug/ml. 2.9 Osteoblast cell culture Osteoblasts isolated from human femoral head trabecular bone (removed during total hip arthroplasty) were cultured
123
Statistical analysis was by ANOVA and Student’s T test between polycaprolactone (as the control) and each pSicontaining material. Standard error of the mean is shown in the figures, with n = 3–6.
3 Results Silicon wafers were electrochemically etched to 70 % porosity and then milled to obtain particle size fractions of 5–35 lm, with a mean average diameter of 11 lm. These were incorporated into a polycaprolactone-acetone solution and the resulting composite formed with a uniform distribution of pSi microparticles (Fig. 1a). The mechanical properties of the composite were evaluated by performing tensile strength testing and shear resistance on polycaprolactone and composites containing 4 and 8 % pSi (Fig. 1b– d). The stress–strain curve of an 8 % pSi–PCL composite, annotated to show the main features of the curve including the linear phase and failure mode is shown in Fig. 1b. The elastic modulus (Fig. 1c) was highest for 8 % composites (89.9 MPa), followed by 4 % composites (79.6 MPa)— both significantly higher than PCL-only (43.4 MPa). Shear resistance was measured by forming the materials in a threaded steel mould and measuring the force required to push out the casting using a metal dowel (Fig. 1d). The force required to push out the polycaprolactone-only
J Mater Sci: Mater Med
material was 1.035 kN and very consistent (±0.01, n = 4). For 4 % composites, the force was 1.001 kN (±0.006) and for 8 % composites 0.996 kN (±0.052), showing only a 3.7 % decrease in shear resistance with the higher level of inclusion, but slightly greater variation between samples (n = 3). Incubation of the composite materials containing between 0.4 and 12 % pSi (weight percentage, w/w) in 500 ml water for 90 days resulted in the formation of soluble silicic acids as a product of pSi dissolution. It was found that the concentration of silicic acid in solution was proportional to the amount of pSi in the composite, with silicic acid concentrations ranging from 5 to 352 ng/ml (Fig. 2a). The bioactivity of pSi was compared to bioactive glass by forming composites made from polycaprolactone with 8 % (w/w) either 45S5 Bioglass particles or pSi and incubation in water for 42 days (Fig. 2c). Both materials yielded silicic acids at a consistent rate (Bioglass,
R2 = 0.859; pSi, R2 = 0.856) and pSi–PCL was shown to release approximately 2.5-fold more silicic acid than Bioglass–PCL at each time point (P \ 0.001). The degradation profiles of the composites were compared to polycaprolactone (Fig. 2a). After 90 days incubation in water, the polycaprolactone-only materials lost 6–8 % of their initial mass due to degradation, whilst those containing pSi either lost less net mass or increased in mass by up to 4 % due to polycondensation of hydrated silica. In simulated body fluid, calcium deposition on composites containing more than 4 % pSi was significantly higher than polycaprolactone-only controls (P \ 0.01) and up to 13-fold higher for composites containing 12 % silicon, showing a clear relationship between pSi content and calcium phosphate deposition (Fig. 2b). Scanning electron microscopy of composites incubated in simulated body fluid for 28 days shows the formation of
A
B
Fig. 2 a The formation of silicic acids in solution and change in mass of the composites over 90 days is due to degradation of polycaprolactone and porous silicon in water. Materials containing porous silicon increased in mass due to the polymerisation and condensation of silicic acids into a hydrated silica gel layer on the surface of the material, corresponding to the increase in silicic acid concentration of the surrounding media. b Calcium deposition on polylactone discs containing 0–12 % porous silicon and incubated in simulated body fluid for 28 days shows that surface mineralisation was relatively low
C
in the silicon-free and low-silicon materials, whereas the inclusion of C4 % porous silicon significantly increases the amount of calcium deposited on the surface of the material compared to the silicon-free control. c Comparison of silicic acid release rates of composites containing either 45S5 Bioglass or porous silicon (8 % w/w in both cases) over 42 days incubation in water shows that pSi-composites released approximately 2.59 more silicic acid at each time point over 6 weeks. Error bars show standard error of the mean, n = 4 (** P \ 0.01; *** P \ 0.001)
123
J Mater Sci: Mater Med
a surface silica gel layer (Fig. 3a–f). Deposits with the morphology of hydroxyapatite formed on all materials including polycaprolactone, but were more prevalent on the composite materials. Further analysis of these deposits using X-ray dispersive spectroscopy shows that the calcium/phosphate ratio was approximately 1.5 and directly proportional to silicon content (R2 = 0.94, Fig. 3g, h). Macrophage morphology was assessed using CLSM as both polycaprolactone and Thermanox have some autofluorescent properties, making conventional fluorescence microscopy difficult. Macrophage activation is characterised by increasing surface membrane ruffles and cellular processes which would be visualised by the actin cytoskeletal staining—this was not observed on any composite, PCL or Thermanox surface after 48 h. The morphology of the nucleus was also similar for all surfaces, indicating that there was no cell death (Fig. 4a). Macrophage activation was measured using an ELISA kit for IL-1b, the production of which was not detected in any PCL, Thermanox or composite sample but was present in the positive controls containing LPS and in the assay internal quality control sample. Macrophage activation was also measured using H2O2 production. Macrophages grown on pSi composites produced very low levels of H2O2 which were not significantly different to those on PCL and were significantly less than cells grown on TCP (P \ 0.022) and the positive controls PMA (P \ 0.025) and 20-fold less than copper (P \ 3 9 10-7) (Fig. 4b). It was noted that whilst 0.5 % pSi slightly increased H2O2 production over polycaprolactone-alone, increasing the pSi concentration proportionally decreased the macrophage response (R2 = 0.976; Fig. 4c). Osteoblast proliferation, as measured by DNA quantification, was rapid on Thermanox (tissue culture plastic) coverslips within the first 7 day s (Fig. 5a). Proliferation on the polycaprolactone-only disc was extremely slow, with no significant increase in cell number over the 28 days. Proliferation on the composite materials was directly proportional to pSi-content, with 0.5 % pSi composites supporting 40 % more osteoblasts than the blank controls and 4.5 % pSi composites supporting 212 % more cells than polycaprolactone alone, although this population was still fourfold lower than supported on Thermanox coverslips. Peak alkaline phosphatase for cells cultured on polycaprolactone and all composites was observed at the 14 day time point, but remained consistently higher on thermanox throughout the experiment, decreasing slightly over time. There was no significant difference in the alkaline phosphatase activity under any of the conditions. Collagen production by the osteoblasts was lowest on polycaprolactone only and increased proportionally with pSi content in the material (Fig. 5b, c). After 28 days, the total amount of collagen on 3 and 4.5 % collagen discs was
123
significantly higher than polycaprolatone-only controls (P \ 0.001) and comparable to Thermanox. Taking into account cell number (collagen/DNA), collagen production per cell appeared to be 3.4-fold higher on 4.5 % pSicomposites than Thermanox (P = 0.315).
4 Discussion In this study we found that composites containing pSi released approximately 2.5 times more silicic acid into solution than composites containing an equal amount of powdered 45S5 Bioglass. The comparatively larger amounts of silicic acid that result from dissolution of pSi suggest potential applications in biomaterials-based approaches for regenerative medicine in instances where bioactive glass composites have proved effective. Previous studies using Bioglass to confer bioactive properties to polymers have relied on substantially greater amounts of the bioactive component, typically 30 [29] to 50 % Bioglass [30] whereas in this study significant effects were achieved from composites containing just 4–12 % pSi. The polycondensation of silicic acids released from pSi composites resulted in the deposition of a silica gel layer on the surface of the composite approximately 1–2 lm thick. In bioactive glass, the silica gel layer formed on the surface of the material acts as a nucleation bed for calcium phosphate [10]. Porosified silicon wafers are also shown to spontaneously nucleate calcium phosphate from SBF on the surface of the wafer [11] and in this study composites incubated in acellular simulated body fluid also formed calcium deposits on the surface of the material. When quantified these were shown to be proportional to the amount of silicic acid released by the material. This calcium phosphate formed on the composite was observed to have the characteristic morphology of biological apatite and the ratio of calcium to phosphorus was proportional to the amount of pSi in the composite, with increasing pSi resulting in increasingly silicate-substituted hydroxyapatite. Silicon-substituted hydroxyapatites are a class of biomaterial that has received considerable attention in recent years due to their enhanced biocompatibility over conventional hydroxyapatites used for tissue engineering (possibly related to their grain size and solubility) [31–33], and so it is interesting that pSi is able to spontaneously generate them in simulated body fluid whilst also yielding bioavailable silicic acids in solution. Bioavailable silicon also plays an important role in the development of the organic phase of bone, having many effects on the extracellular matrix by helping to stabilise the glycosaminoglycan network via silanolate bonds and stimulating upregulation of collagen I expression by osteoblasts [3, 4, 34].
J Mater Sci: Mater Med
A
B
C
D
E
F
G
H
Fig. 3 eSEM images showing hydroxyapatite formation on a polycaprolactone and composites of polycaprolactone containing b 0.4 %, c 4 % and d 8 % porous silicon microparticles. In b–d a cracked silica gel layer is also visible, a cross-section of which is visible in e (0.4 % pSi) showing it to be 1–2 lm thick. Some 4 % composites showed substantial mineralisation (f). Calcium:phosphate ratios of
hydroxyapatite deposited on the surface of polycaprolactone and polycaprolactone–pSi composite discs as determined energy-dispersive X-ray spectroscopy. g The Ca:P ratio is proportional to the amount of porous silicon present in the material (h, R2 = 0.939). Error bars show standard error of the mean, n = 3
123
J Mater Sci: Mater Med
A
B
C
Fig. 4 Response of J774.A1 murine macrophage cell line to porous silicon-polycaprolactone composites. a morphology of macrophages incubated for 48 h on pSi–PCL composite was similar to that in either tissue culture plastic or PCL polymer controls. No excess cell ruffling or increased numbers of cellular processes were seen, indicating that cells were not activated. b Hydrogen peroxide production by murine macrophages was measured as an indicator of cell activation after 6 h. Copper surfaces (Cu) and phorbol 12-myristate 13-acetate (PMA) surfaces were used as positive controls, as known substrates causing
macrophage activation. Tissue culture plastic (TCP) was also used as a control to which all data were compared using a two-way ANOVA. Hydrogen peroxide production from macrophages grown on any pSi– PCL composites was not significantly different to those grown on PCL but significantly lower than those on TCP, PMA or copper, whilst increasing pSi content in the composite proportionally decreased the macrophage response in terms of H2O2 production (c). (* P \ 0.05)
Macrophage studies clearly demonstrated that the pSi–PCL composites do not elicit a detrimental immune response. Morphological evidence of macrophage activation such as increased numbers of cell processes and surface membrane ruffling [35] was absent and biochemical measurements of hydrogen peroxide and IL-1b further demonstrated the lack of response. In fact, no material caused macrophages growing on the surface to secrete IL-1b (an important inflammatory cytokine) though the macrophages were able to produce IL-1b in dose dependant quantities when stimulated with LPS. Hydrogen peroxide
production remained low for macrophages cultured on tissue culture plastic, PCL and pSi–PCL composites and was significantly increased over 6 h for macrophages cultured on copper and PMA, surfaces known to elicit a macrophage response [36]. Macrophages actually produced less hydrogen peroxide on pSi–PCL composites than on tissue culture plastic, in a trend suggesting that increasing pSi content correlates with decreasing hydrogen peroxide production. Culturing primary human osteoblasts on the surface of polycaprolactone, pSi-composite discs and Thermanox
123
J Mater Sci: Mater Med
A
B
C
Fig. 5 The activity of human osteoblasts on the surface of polycaprolactone (PCL), pSi–PCL composites and thermanox coverslips (tissue culture plastic) was compared. a After 28 days, the DNA content of lysates from cells cultured on PCL-based materials was lower than Thermanox controls, but proportional to the amount of silicon in the composite (gradient:182 ng DNA/mg pSi). b The production of collagen by osteoblasts was again highest on
Thermanox, and lowest on the polycaprolactone-only disk; increasing pSi content in the composite resulted in an increased rate of collagen production. c After 28 days, the amount of collagen produced on composites containing 3 or 4.5 % pSi was significantly higher than polycaprolactone-only and equivalent to the amount produced on Thermanox. Error bars show standard error of the mean, n = 6 (composites), n = 12 (Thermanox), (*** P \ 0.001)
revealed that whilst Thermanox supports rapid cell growth in monolayer, the osteoblastic phenotype is less pronounced in terms of alkaline phosphatase expression and collagen production (when normalised for cell number). Conversely, polycaprolactone alone is a poor substrate for cell growth, but the addition of increasing quantities of pSi to the material resulted in proportional increases in cell proliferation, as seen with some other silicon-containing biomaterials [37]. Alkaline phosphatase expression steadily decreased over 28 days when cells were cultured on Thermanox, but exhibited peak activity at 14 days on the composites, with a subsequent decrease to a lower level of constitutive expression. This general trend in expression has been observed and reported by many biomaterials research groups, particularly those using primary osteoblasts [38, 39] and silicate/bioglass-based materials [40] and is also found in vivo [41]. Osteoblast production of collagen was also increased with the addition of pSi and, when normalised for cell number, collagen production was significantly higher on the pSi composite than Thermanox, indicating that the mechanism of silicon-induced osteoblast enhancement of osteogenesis
is via increased matrix production and cellular proliferation [37, 40, 41].
5 Conclusion In this work, the bioactivity of pSi—polycaprolactone composites has been characterised by a variety of means. Incorporating pSi with the polymer enhanced the deposition of hydroxyapatite on the surface of the composite over polycaprolactone controls, and this enhancement was found to be proportional to increasing silicon content. Furthermore, silicon-substituted hydroxyapatites appear to spontaneously form on the composites in simulated body fluid. Adding up to 8 per cent pSi to polycaprolactone increased the elastic modulus of the material and did not have any significant negative impact on the mechanical properties. At these concentrations, pSi supported osteoblast growth and appeared to reduce the response of cultured macrophages in vitro. For this initial investigation into polymer composites containing pSi, the biomaterial has been demonstrated to
123
J Mater Sci: Mater Med
possess similar characteristics to bioactive glass: forming silicic acids in solution which then polycondense to form surface layers of silica and apatite. It is the novel features of pSi which make it most interesting to the biomaterials community: having an open porous structure which can be surface modified, filled with pharmaceuticals and possessing optoelectronic properties derived from quantum phenomena. As silicon technology is so ubiquitous within the modern world, the development of pSi as a biomaterial may lead to fascinating interdisciplinary applications and provide further insights into the role of silicon in the biology of bone.
References 1. Carlisle EM. Silicon as an essential element. Fed Proc. 1974;33:1758–66. 2. Carlisle EM. Silicon as a trace nutrient. Sci Total Environ. 1988;73:95–106. 3. Reffitt DM, Ogston N, Jugdaohsingh R, Cheung HFJ, Evans BAJ, Thompson RPH, Powell JJ, Hampson GN. Orthosilicic acid stimulates collagen type 1 synthesis and osteoblastic differentiation in human osteoblast-like cells in vitro. Bone. 2003;32: 127–35. 4. Arumugam MQ, Ireland DC, Brooks RA, Rushton N, Bonfield W. Orthosilicic acid increases collagen type I mRNA expression in human bone-derived osteoblasts in vitro. Key Eng Mater. 2006;254:869–72. 5. Anderson SI, Downes S, Perry CC, Caballero AM. Evaluation of the osteoblast response to a silica gel in vitro. J Mater Sci Mater Med. 1998;9:731–5. 6. Xynos IE, Alasdair JE, Buttery LDK, Hench LL, Polak JM. Gene-expression profiling of human osteoblasts following treatment with the ionic products of Bioglass 45S5. J Biomed Mater Res. 2001;55:151–7. 7. Valerio P, Pereira MM, Goes AM, Leite MF. The effect of ionic products from bioactive glass dissolution on osteoblast proliferation and collagen production. Biomaterials. 2004;25:2941–8. 8. Sripanyakorn S, Jugdaohsingh R, Thompson RPH, Powell JJ. Dietary silicon and bone health. Nutr Bull. 2005;30:222–30. 9. Jugdaohsingh R, Tucker KL, Qiao N, Cupples LA, Kiel DP, Powell JJ. Dietary silicon intake is positively associated with bone mineral density in men and premenopausal women of the Framingham Offspring cohort. J Bone Miner Res. 2004;19: 297–307. 10. Hench LL. The story of BioglassÒ. J Mater Sci. 2006;17:967–78. 11. Canham T. Bioactive silicon fabrication via nanoetching techniques. Adv Mater. 1995;7:1033–7. 12. Anglin EJ, Cheng L, Freeman WR, Sailor MJ. Porous silicon in drug delivery devices and materials. Adv Drug Deliv Rev. 2008;60:1266–77. 13. Serda RE, Gu J, Bhavane RC, Liu X, Chiappini C, Decuzzi P, Ferrari M. The association of silicon microparticles with endothelial cells in drug delivery to the vasculature. Biomaterials. 2009;30:2440–8. 14. Zhang K, Loong SL, Connor S, Yu SW, Tan SY, Ng RT, Lee KM, Canham L, Chow PK. Complete tumor response following intratumoral 32P BioSilicon on human hepatocellular and pancreatic carcinoma xenografts in nude mice. Clin Cancer Res. 2005;11:7532–7.
123
15. Park JH, Gui L, Malzahn G, Ruoslahti E, Bhatia SN, Sailor MJ. Biodegradable porous silicon nanoparticles for in vivo applications. Nat Mater. 2009;8:331–6. 16. Coffer JL, Montchamp JL, Aimone JB, Weis RP. Routes to calcifled porous silicon: implications for drug delivery and biosensing. Physica Status Solidi A. 2003;197:336–9. 17. Anglin EJ, Schwartz MP, Ng VP, Perelman LA, Sailor MJ. Engineering the chemistry and nanostructure of porous silicon Fabry–Pe´rot films for loading and release of a steroid. Langmuir. 2004;20:11264–9. 18. Charnay C, Begu S, Tourne-Peteilh C, Nicole L, Lerner DA, Devoisselle JM. Inclusion of ibuprofen in mesoporous templated silica: drug loading and release property. Eur J Pharm Biopharm. 2004;57:533–40. 19. Vaccari L, Canton D, Zaffaroni N, Villa R, Tormen M, di Fabrizio E. Porous silicon as drug carrier for controlled delivery of doxorubicin anticancer agent. Microelectron Eng. 2006;83: 1598–601. 20. Salonen J, Kaukonen AM, Hirvonen J, Lehto V-P. Mesoporous silicon in drug delivery applications. J Pharm Sci. 2008;97:632–53. 21. Whitehead MA, Fan D, Mukherjee P, Akkaraju GR, Canham LT, Coffer JL. High-porosity poly(epsilon-caprolactone)/mesoporous silicon scaffolds: calcium phosphate deposition and biological response to bone precursor cells. Tissue Eng Part A. 2008; 14:195–206. 22. Johansson F, Kanje M, Linsmeier CE, Wallman L. The influence of porous silicon on axonal outgrowth in vitro. IEEE Trans Biomed Eng. 2008;55:1447–9. 23. Alvarez SD, Derfus AM, Schwartz MP, Bhatia SN, Sailor MJ. The compatibility of hepatocytes with chemically modified porous silicon with reference to in vitro biosensors. Biomaterials. 2008;30:26–34. 24. Dash TK, Konkimalla VB. Poly-E-caprolactone based formulations for drug delivery and tissue engineering: a review. J Control Release. 2012;158:15–33. 25. Sinha VR, Trehan A. Development, characterization, and evaluation of ketorolac tromethamine-loaded biodegradable microspheres as a depot system for parenteral delivery. Drug Deliv. 2008;15:365–72. 26. Rohner D, Hutmacher DW, Cheng TK, Oberholzer M, Hammer B. In vivo efficacy of bone-marrow-coated polycaprolactone scaffolds for the reconstruction of orbital defects in the pig. J Biomed Mater Res B. 2003;66:574–80. 27. Halimaoui A. Porous silicon formation by anodization. In: Canham LT, editor. Properties of porous silicon. London: Institution of Engineering and Technology; 1997. 28. Kokubo T, Kushitani H, Sakka S, Kitsugi T, Yamamuro T. Solutions able to reproduce in vivo surface-structure changes in bioactive glass-ceramic A-W. Biomed Mater Res. 1990;24:721–34. 29. Chouzouri G, Xanthos M. In vitro bioactivity and degradation of polycaprolactone composites containing silicate fillers. Acta Biomater. 2007;3:745–56. 30. Venugopal JR, Low S, Choon AT, Kumar AB, Ramakrishna S. Nanobioengineered electrospun composite nanofibers and osteoblasts for bone regeneration. Artif Organs. 2008;32:388–97. 31. Porter AE, Patel N, Skepper JN, Best SM, Bonfield W. Comparison of in vivo dissolution processes in hydroxyapatite and silicon-substituted hydroxyapatite bioceramics. Biomaterials. 2003;24:4609–20. 32. Porter AE, Patel N, Skepper JN, Best SM, Bonfield W. Effect of sintered silicate-substituted hydroxyapatite on remodelling processes at the bone-implant interface. Biomaterials. 2004;25: 3303–14. 33. Gibson IR, Best SM, Bonfield W. Effect of silicon substitution on the sintering and microstructure of hydroxyapatite. J Am Ceram Soc. 2002;85:2771–7.
J Mater Sci: Mater Med 34. Schwarz K. A bound form of silicon in glycosaminoglycans and polyuronides. Proc Natl Acad Sci. 1973;70: 1608–12. 35. Kleveta G, Borze˛cka K, Zdioruk M, Czerkies M, Kuberczyk H, Sybirna N, et al. LPS induces phosphorylation of actin-regulatory proteins leading to actin reassembly and macrophage motility. J Cell Biochem. 2012;113:80–92. 36. Scotchford CA, Garle MJ, Batchelor J, Bradley J, Grant DM. Use of a novel carbon fibre composite material for the femoral stem component of a THR system: in vitro biological assessment. Biomaterials. 2003;24:4871–9. 37. Keeting PE, Oursler MJ, Weigand KE, Bonde SK, Spelsberg TC, Riggs BL. Zeolite A increases proliferation, differentiation and transforming growth factor beta production in normal adult human osteoblast-like cells in vitro. J Bone Miner Res. 1992;7: 1281–9.
38. Sammons RL, El Haj AJ, Marquis PM. Novel culture procedure permitting the synthesis of proteins by rat calvarial cells cultured on hydroxyapatite particles to be quantified. Biomaterials. 1994;15:536–42. 39. Homaeigohar SSh, Shokrgozar MA, Sadi AY, Khavandi A, Javadpour J, Hosseinalipour M. In vitro evaluation of biocompatibility of beta-tricalcium phosphate-reinforced high-density polyethylene; an orthopedic composite. J Biomed Mater Res A. 2005;75:14–22. 40. Kubo K, Tsukasa N, Uehara M, Izumi Y, Ogino M, Kitano M, Sueda T. Calcium and silicon from bioactive glass concerned with formation of nodules in periodontal-ligament fibroblasts in vitro. J Oral Rehabil. 1997;24:70–5. 41. Zhou H, Choong P, McCarthy R, Chou ST, Martin TJ, Ng KW. In situ hybridization to show sequential expression of osteoblast gene markers during bone formation in vivo. J Bone Miner Res. 1994;9:1489–99.
123
Porous silicon confers bioactivity to polycaprolactone composites in vitro J. R. Henstock • U. R. Ruktanonchai L. T. Canham • S. I. Anderson
•
Received: 19 September 2013 / Accepted: 2 January 2014 Ó Springer Science+Business Media New York 2014
Abstract Silicon is an essential element for healthy bone development and supplementation with its bioavailable form (silicic acid) leads to enhancement of osteogenesis both in vivo and in vitro. Porous silicon (pSi) is a novel material with emerging applications in opto-electronics and drug delivery which dissolves to yield silicic acid as the sole degradation product, allowing the specific importance of soluble silicates for biomaterials to be investigated in isolation without the elution of other ionic species. Using polycaprolactone as a bioresorbable carrier for porous silicon microparticles, we found that composites containing pSi yielded more than twice the amount of bioavailable silicic acid than composites containing the same mass of 45S5 Bioglass. When incubated in a simulated body fluid, the addition of pSi to polycaprolactone significantly increased the deposition of calcium phosphate. Interestingly, the apatites formed had a Ca:P ratio directly proportional to the silicic acid concentration, indicating that silicon-substituted hydroxyapatites were being spontaneously formed as a first J. R. Henstock (&) Institute for Science and Technology in Medicine, Keele University, Stoke-on-Trent ST4 7QB, UK e-mail: [email protected] U. R. Ruktanonchai National Nanotechnology Center (NANOTEC), National Science and Technology Development Agency (NSTDA), Thailand Science Park, Pathumthani 12120, Thailand L. T. Canham pSiMedica Ltd., Malvern Hills Science Park, Geraldine Road, Malvern, Worcestershire WR14 3SZ, UK S. I. Anderson Royal Derby Hospital Centre, School of Graduate Entry Medicine and Health, University of Nottingham, Uttoxeter Road, Derby DE22 3DT, UK
order reaction. Primary human osteoblasts cultured on the surface of the composite exhibited peak alkaline phosphatase activity at day 14, with a proportional relationship between pSi content and both osteoblast proliferation and collagen production over 4 weeks. Culturing the composite with J744A.1 murine macrophages demonstrated that porous silicon does not elicit an immune response and may even inhibit it. Porous silicon may therefore be an important next generation biomaterial with unique properties for applications in orthopaedic tissue engineering.
1 Introduction Silicon is an essential element in human nutrition, with the symptoms of a silicon-deficient diet including abnormal bone development [1–4]. Similarly, animals and bone forming osteoblast-like cells in vitro show an increase in osteogenesis when supplemented orthosilicic acid [5–7] the only bioavailable source of silicon for higher animals [8]. Research has shown that supplementing dietary silicon content in humans is associated with improved bone mineral density (BMD), with supplementation of 40 mg/day increasing cortical hip BMD by up to 10 % [9]. Silicic acids are released from Bioglass and confer bioactivity by polycondensing in vivo to form a hydrated silica gel layer which adsorbs growth factors, supports cell growth and acts a nucleation bed for bone mineral, facilitating a strong chemical bond to living bone [10]. Porous silicon (pSi) exhibits several of the same characteristics as bioactive glass by forming a surface layer of hydroxyapatite in simulated body fluid resulting from the release of silicic acids in aqueous solution and is therefore of interest as a potential bioactive material [11]. pSi is made by the electrochemical etching of polycrystalline wafers, creating
123
J Mater Sci: Mater Med
nanoscale silicon hydride-lined pores through the material and resulting in an aqueously soluble material with a large surface area to volume ratio (500 m2/cm3) and hence high reactivity [11]. Porous silicon has many potential uses in the fields of opto-electronics and an emerging number of biological and clinical applications. The long, narrow pores can be filled with pharmaceuticals, creating a dissolvable drug delivery material with release kinetics that are easily controllable by adjusting the pore morphology and drug loading density [12, 13]. pSi is also being studied for applications in radioisotope intratumoral brachytherapy for treatment of cancers such as pancreatic and hepatocellular carcinoma [14]. It has been shown that injected pSi particles are fully degraded into silicic acid and cleared by excretion via the kidneys [15]. Drug delivery using dispersible pSi particles has been explored by several research groups with the release kinetics of various molecules including cis-platin [16], dexamethasone [12, 17], ibuprofen [18], and doxorubicin [19] having been reported. A comprehensive review of pSi for drug delivery applications is given in Salonen et al. [20]. Porosified silicon wafers have been shown to have low cytotoxicity, supporting the growth in vitro of several cell types including osteoblasts [21], neurones [22], and hepatocytes [23]. However, like bioactive glasses, the mechanical properties of high porosity silicon are poor, and so for these experiments we developed a composite material to explore the bioactive properties of pSi within a strong but resorbable polycaprolactone matrix, a polymer which has existing uses in tissue engineering [24], drug delivery [25], and is FDA approved for human in vivo use [26]. In this investigation, we have evaluated a range of polycaprolactone composites containing 0.4–12 % pSi by weight and characterised them by their formation of silicic acids, silica gel and hydroxyapatite in simulated body fluid. The aims of this research are initially to assess pSi as a bioactive material with unique properties suitable for inclusion in polymer-based composites for orthopaedic applications.
2 Materials and methods 2.1 Porous silicon Porous silicon for this investigation was produced by the anodisation of an electronic-grade silicon wafer (200 mm diameter, 725 lm thick) using a platinum cathode and immersed in a hydrogen fluoride electrolyte. A direct current is passed through the cell which corrodes the anode wafer, producing an even, porous layer which was dried, milled and filtered to produce 5–35 lm pSi particles with
123
an average diameter of 11 lm (Fig. 1a). A comprehensive methodology for silicon porosification using this method is given in Halimaoui [27]. 2.2 Composite formation A composite of polycaprolactone and pSi microparticles was made by dissolving the polymer in a small volume of acetone at 40 °C and mixing in pSi particles in weight percentages of 0.4–12 %. The acetone was evaporated at room temperature in a flow hood to create a solid composite which was cut into sections and placed in a 5 ml polypropylene syringe body. After heating to 70 °C to melt together the composite the cylinder was cooled until solid, removed from the syringe body and cut into 10 mm diameter 250 mg discs containing 1 (0.4 % w/w) to 30 mg (12 % w/w) silicon which were used for subsequent testing and compared to polycaprolactone discs made in the same way without pSi. To visualise the particle distribution, 1 lg/ml fluorecein solution was adsorbed into to the pSi particles prior to inclusion by brief immersion followed by filtration. 2.3 Mechanical testing The tensile strength of the composite was tested using a cylinder of the material cast in a 1 ml polypropylene syringe body. Polycaprolactone casts were tested against 4 and 8 % pSi–PCL composites. The samples were fitted into a 50 KN load cell in an Instron 5569 Tensile Testing System to test a 20 mm region of the sample. Samples were fitted into grips which were moved apart at 2.0 mm/min until fail. Maximum load and tensile stress at maximum load were measured. Screw-shaped composites were produced by melting the material inside a stainless steel threaded mould (internal diameter 12 mm, length 14 mm) on a hotplate at 80 °C and then cooling the cast inside the mould until solid. The steel mould was held in a recessed block over a central hole and a 6 mm diameter steel dowel clamped in the upper grip of the Instron testing system. The dowel was lowered to touch the polymer/composite surface and then pushed out through the mould at 2 mm/min. The force (N) required to push out the casting was measured for PCL, 4 and 8 % composites and the failure mode determined by SEM of sections through the pushed-out material. 2.4 Bioactivity assays The primary test for bioactivity involved immersing the composite in simulated body fluid (SBF), prepared as described by Kokubo et al. [28]. Composite discs were incubated for periods of 28–90 days and compared to polycaprolactone as a control. pSi was compared to
J Mater Sci: Mater Med
A
B
C
D
Fig. 1 a The 70 % porous silicon microparticles used in this investigation, stained with fluorescein and imaged using a confocal microscope showing the even distribution of 11 lm particles within the polycaprolactone composite. b Typical stress–strain curve of composite (8 % pSi) showing the locations on the curve of the proportionality limit beyond which stress/strain is non-linear and
below which Young’s modulus is calculated, the ultimate (or yield) stress, and the beginning and mode of failure. c Elastic modulus was calculated from the gradient of the stress–strain curve below the elastic limit. d The force required to push out the casting from the mold using a metal dowel. Error bars are standard error of the mean, n = 4. Scale bar 100 lm
bioactive glass by forming composites made from polycaprolactone with 8 % (w/w) either 45S5 Bioglass particles or pSi and incubating in water for 42 days.
weight before incubation and after 90 days, compared to polycaprolactone-only controls. Confirmation of silicia gel and apatite formation was by scanning electron microscopy and X-ray microanalysis (EDX) in samples imaged using a Phillips XL30 FEG-ESEM. After incubation for 28 days in SBF, the discs were rinsed, dried and imaged at 10 kV in auxiliary mode with nitrogen as the column gas allowing energy dispersive X-ray microanalysis to be performed on uncoated samples. The elemental composition of surface layers was mapped using the integrated elemental mapping software. Additional SEM (e.g. on castings for mechanical testing) was performed on samples which were gold coated and imaged using a Jeol JSM840 scanning electron microscope.
2.5 Silicon degradation Analysis of silicic acid in the above solutions (resulting from the degradation of pSi) was by a molybdenum (molybdate) blue assay, in which ammonium molybdate in solution reacts with silicic acid in the sample to form bsilicomolybdic acid (yellow) which is subsequently reduced to form a molybdate blue complex with a peak absorbance at 810 nm. Experimental values for soluble silicon species (ortho- and polysilicic acids) were calibrated against known concentrations of sodium metasilicate.
2.7 Macrophage cell culture 2.6 Silica gel and apatite formation Polycondensation of silica on the material was evaluated by incubating the composites in dH2O and comparing dry
The murine BALB/c monocyte macrophage cell line J774A.1 was used to determine the effect of the composite on macrophage morphology and activation. Composite
123
J Mater Sci: Mater Med
samples of PCL containing 0.5, 1.5, 3 or 4.5 % pSi (equating to 1, 3, 6 or 9 mg/200 mg composite disc) were placed in the wells of a 24 well plate (Nunc). Thermanox tissue culture plastic coverslips and PCL polymer samples were used as controls. Cells were seeded at a density of 2 9 105 cells/ml in Dulbecco’s Modified Eagles Medium (DMEM) containing 10 % FCS, 5 % sodium bicarbonate, 1 % glutamine and 1 % penicillin/streptomycin and incubated at 37 °C in a humidified atmosphere containing 5 % CO2 for 48 h. After 48 h, samples were washed in warmed sterile PBS and fixed in 4 % paraformaldehyde for 5 min. Samples were permeabilised in a solution containing triton X100 for 10 min at 0C and the f-actin was stained with FITC-conjugated phalloidin (Sigma) for 20 min. The cell nuclei were then stained with PI for 30 s. Samples were cover-slipped and viewed in a Leica TCS-4D confocal laser scanning microscope (CLSM) using the 488 nm laser line to excite the green fluorescence of FITC-conjugated phalloidin and the 568 nm laser line to excite the red fluorescence of PI. Optical sections were taken at 0.3 mm intervals and displayed as a maximum intensity projection using Scanware software.
in DMEM supplemented with 10 % FCS, 2 mM L-glutamine, 1 % non-essential amino acids, 0.2 mM HEPES, 150 lg/ml ascorbic acid, 2 % penicillin/streptomycin and with 10 mM sodium b-glycerophosphate and 10-7 M dexamethasone (all Sigma, UK). Osteoblasts were seeded at a density of 80,000 cells/ml directly onto the polycaprolactone or composite disc surface or thermanox coverslip (Nunc) in 24-well plates (Nunc). Media was replaced every 48 h and the osteoblasts incubated at 37 °C in 5 % CO2. Cell lysates (n = 6) were obtained at intervals of 7, 14, 21 and 28 days by freezing the monolayer cells in dH2O. DNA was quantified by incubating cell lysate samples with 1 % bisBenzimide H (Hoechst 33258) dye (Sigma, UK). Samples were measured and the fluorescence was read at excitation 355 nm and emission 460 nm using a Fluoroskan Ascent reader and compared to a standard curve of calf thymus DNA (Sigma, UK). Alkaline phosphatase (ALP) was quantified using p-nitrophenyl phosphate (Sigma, UK) which forms a complex with ALP present in the cell lysate with a peak absorbance read at 405 nm. Collagen was quantified in cell lysates using the sirius red-based Sircol collagen assay kit (Biocolor, UK) at 540 nm and compared to a standard curve of type-I rat tail collagen.
2.8 Macrophage activation 2.10 Statistics Macrophage activation was quantified using the Quantikine M Mouse IL-1b immunoassay (ELISA) (R&D Systems) to determine the levels of IL-1b produced and secreted into the culture medium by macrophages growing on the different surfaces. Samples were prepared and cells seeded as above. Macrophages grown on tissue culture plastic (TCP) were used as the negative control and macrophages grown on TCP containing 10 or 100 mg/ml lipopolysaccharide (LPS), a known macrophage activator, were used as the positive control. After 48 h the medium was removed from the wells and the assayed with the absorbance measured at 450 nm for test samples and known concentrations of IL1b. Quantities of IL-1b were determined by interpolation from the standard curve. The production of hydrogen peroxide by J774A.1 macrophages seeded on composite surfaces was measured using dichlorofluorescein (DCF). Cells were seeded at a concentration of 2 9 105 cells/ml as before. TCP and PCL were used as controls. Copper substrates were also used as positive controls. Cells were incubated in DCF (1:100 in Hank’s Balanced Salt Solution). PMA (Phorbol 12-Myristate 13-Acetate, Sigma) was used as a positive control and was added to the appropriate wells at 15 ug/ml. 2.9 Osteoblast cell culture Osteoblasts isolated from human femoral head trabecular bone (removed during total hip arthroplasty) were cultured
123
Statistical analysis was by ANOVA and Student’s T test between polycaprolactone (as the control) and each pSicontaining material. Standard error of the mean is shown in the figures, with n = 3–6.
3 Results Silicon wafers were electrochemically etched to 70 % porosity and then milled to obtain particle size fractions of 5–35 lm, with a mean average diameter of 11 lm. These were incorporated into a polycaprolactone-acetone solution and the resulting composite formed with a uniform distribution of pSi microparticles (Fig. 1a). The mechanical properties of the composite were evaluated by performing tensile strength testing and shear resistance on polycaprolactone and composites containing 4 and 8 % pSi (Fig. 1b– d). The stress–strain curve of an 8 % pSi–PCL composite, annotated to show the main features of the curve including the linear phase and failure mode is shown in Fig. 1b. The elastic modulus (Fig. 1c) was highest for 8 % composites (89.9 MPa), followed by 4 % composites (79.6 MPa)— both significantly higher than PCL-only (43.4 MPa). Shear resistance was measured by forming the materials in a threaded steel mould and measuring the force required to push out the casting using a metal dowel (Fig. 1d). The force required to push out the polycaprolactone-only
J Mater Sci: Mater Med
material was 1.035 kN and very consistent (±0.01, n = 4). For 4 % composites, the force was 1.001 kN (±0.006) and for 8 % composites 0.996 kN (±0.052), showing only a 3.7 % decrease in shear resistance with the higher level of inclusion, but slightly greater variation between samples (n = 3). Incubation of the composite materials containing between 0.4 and 12 % pSi (weight percentage, w/w) in 500 ml water for 90 days resulted in the formation of soluble silicic acids as a product of pSi dissolution. It was found that the concentration of silicic acid in solution was proportional to the amount of pSi in the composite, with silicic acid concentrations ranging from 5 to 352 ng/ml (Fig. 2a). The bioactivity of pSi was compared to bioactive glass by forming composites made from polycaprolactone with 8 % (w/w) either 45S5 Bioglass particles or pSi and incubation in water for 42 days (Fig. 2c). Both materials yielded silicic acids at a consistent rate (Bioglass,
R2 = 0.859; pSi, R2 = 0.856) and pSi–PCL was shown to release approximately 2.5-fold more silicic acid than Bioglass–PCL at each time point (P \ 0.001). The degradation profiles of the composites were compared to polycaprolactone (Fig. 2a). After 90 days incubation in water, the polycaprolactone-only materials lost 6–8 % of their initial mass due to degradation, whilst those containing pSi either lost less net mass or increased in mass by up to 4 % due to polycondensation of hydrated silica. In simulated body fluid, calcium deposition on composites containing more than 4 % pSi was significantly higher than polycaprolactone-only controls (P \ 0.01) and up to 13-fold higher for composites containing 12 % silicon, showing a clear relationship between pSi content and calcium phosphate deposition (Fig. 2b). Scanning electron microscopy of composites incubated in simulated body fluid for 28 days shows the formation of
A
B
Fig. 2 a The formation of silicic acids in solution and change in mass of the composites over 90 days is due to degradation of polycaprolactone and porous silicon in water. Materials containing porous silicon increased in mass due to the polymerisation and condensation of silicic acids into a hydrated silica gel layer on the surface of the material, corresponding to the increase in silicic acid concentration of the surrounding media. b Calcium deposition on polylactone discs containing 0–12 % porous silicon and incubated in simulated body fluid for 28 days shows that surface mineralisation was relatively low
C
in the silicon-free and low-silicon materials, whereas the inclusion of C4 % porous silicon significantly increases the amount of calcium deposited on the surface of the material compared to the silicon-free control. c Comparison of silicic acid release rates of composites containing either 45S5 Bioglass or porous silicon (8 % w/w in both cases) over 42 days incubation in water shows that pSi-composites released approximately 2.59 more silicic acid at each time point over 6 weeks. Error bars show standard error of the mean, n = 4 (** P \ 0.01; *** P \ 0.001)
123
J Mater Sci: Mater Med
a surface silica gel layer (Fig. 3a–f). Deposits with the morphology of hydroxyapatite formed on all materials including polycaprolactone, but were more prevalent on the composite materials. Further analysis of these deposits using X-ray dispersive spectroscopy shows that the calcium/phosphate ratio was approximately 1.5 and directly proportional to silicon content (R2 = 0.94, Fig. 3g, h). Macrophage morphology was assessed using CLSM as both polycaprolactone and Thermanox have some autofluorescent properties, making conventional fluorescence microscopy difficult. Macrophage activation is characterised by increasing surface membrane ruffles and cellular processes which would be visualised by the actin cytoskeletal staining—this was not observed on any composite, PCL or Thermanox surface after 48 h. The morphology of the nucleus was also similar for all surfaces, indicating that there was no cell death (Fig. 4a). Macrophage activation was measured using an ELISA kit for IL-1b, the production of which was not detected in any PCL, Thermanox or composite sample but was present in the positive controls containing LPS and in the assay internal quality control sample. Macrophage activation was also measured using H2O2 production. Macrophages grown on pSi composites produced very low levels of H2O2 which were not significantly different to those on PCL and were significantly less than cells grown on TCP (P \ 0.022) and the positive controls PMA (P \ 0.025) and 20-fold less than copper (P \ 3 9 10-7) (Fig. 4b). It was noted that whilst 0.5 % pSi slightly increased H2O2 production over polycaprolactone-alone, increasing the pSi concentration proportionally decreased the macrophage response (R2 = 0.976; Fig. 4c). Osteoblast proliferation, as measured by DNA quantification, was rapid on Thermanox (tissue culture plastic) coverslips within the first 7 day s (Fig. 5a). Proliferation on the polycaprolactone-only disc was extremely slow, with no significant increase in cell number over the 28 days. Proliferation on the composite materials was directly proportional to pSi-content, with 0.5 % pSi composites supporting 40 % more osteoblasts than the blank controls and 4.5 % pSi composites supporting 212 % more cells than polycaprolactone alone, although this population was still fourfold lower than supported on Thermanox coverslips. Peak alkaline phosphatase for cells cultured on polycaprolactone and all composites was observed at the 14 day time point, but remained consistently higher on thermanox throughout the experiment, decreasing slightly over time. There was no significant difference in the alkaline phosphatase activity under any of the conditions. Collagen production by the osteoblasts was lowest on polycaprolactone only and increased proportionally with pSi content in the material (Fig. 5b, c). After 28 days, the total amount of collagen on 3 and 4.5 % collagen discs was
123
significantly higher than polycaprolatone-only controls (P \ 0.001) and comparable to Thermanox. Taking into account cell number (collagen/DNA), collagen production per cell appeared to be 3.4-fold higher on 4.5 % pSicomposites than Thermanox (P = 0.315).
4 Discussion In this study we found that composites containing pSi released approximately 2.5 times more silicic acid into solution than composites containing an equal amount of powdered 45S5 Bioglass. The comparatively larger amounts of silicic acid that result from dissolution of pSi suggest potential applications in biomaterials-based approaches for regenerative medicine in instances where bioactive glass composites have proved effective. Previous studies using Bioglass to confer bioactive properties to polymers have relied on substantially greater amounts of the bioactive component, typically 30 [29] to 50 % Bioglass [30] whereas in this study significant effects were achieved from composites containing just 4–12 % pSi. The polycondensation of silicic acids released from pSi composites resulted in the deposition of a silica gel layer on the surface of the composite approximately 1–2 lm thick. In bioactive glass, the silica gel layer formed on the surface of the material acts as a nucleation bed for calcium phosphate [10]. Porosified silicon wafers are also shown to spontaneously nucleate calcium phosphate from SBF on the surface of the wafer [11] and in this study composites incubated in acellular simulated body fluid also formed calcium deposits on the surface of the material. When quantified these were shown to be proportional to the amount of silicic acid released by the material. This calcium phosphate formed on the composite was observed to have the characteristic morphology of biological apatite and the ratio of calcium to phosphorus was proportional to the amount of pSi in the composite, with increasing pSi resulting in increasingly silicate-substituted hydroxyapatite. Silicon-substituted hydroxyapatites are a class of biomaterial that has received considerable attention in recent years due to their enhanced biocompatibility over conventional hydroxyapatites used for tissue engineering (possibly related to their grain size and solubility) [31–33], and so it is interesting that pSi is able to spontaneously generate them in simulated body fluid whilst also yielding bioavailable silicic acids in solution. Bioavailable silicon also plays an important role in the development of the organic phase of bone, having many effects on the extracellular matrix by helping to stabilise the glycosaminoglycan network via silanolate bonds and stimulating upregulation of collagen I expression by osteoblasts [3, 4, 34].
J Mater Sci: Mater Med
A
B
C
D
E
F
G
H
Fig. 3 eSEM images showing hydroxyapatite formation on a polycaprolactone and composites of polycaprolactone containing b 0.4 %, c 4 % and d 8 % porous silicon microparticles. In b–d a cracked silica gel layer is also visible, a cross-section of which is visible in e (0.4 % pSi) showing it to be 1–2 lm thick. Some 4 % composites showed substantial mineralisation (f). Calcium:phosphate ratios of
hydroxyapatite deposited on the surface of polycaprolactone and polycaprolactone–pSi composite discs as determined energy-dispersive X-ray spectroscopy. g The Ca:P ratio is proportional to the amount of porous silicon present in the material (h, R2 = 0.939). Error bars show standard error of the mean, n = 3
123
J Mater Sci: Mater Med
A
B
C
Fig. 4 Response of J774.A1 murine macrophage cell line to porous silicon-polycaprolactone composites. a morphology of macrophages incubated for 48 h on pSi–PCL composite was similar to that in either tissue culture plastic or PCL polymer controls. No excess cell ruffling or increased numbers of cellular processes were seen, indicating that cells were not activated. b Hydrogen peroxide production by murine macrophages was measured as an indicator of cell activation after 6 h. Copper surfaces (Cu) and phorbol 12-myristate 13-acetate (PMA) surfaces were used as positive controls, as known substrates causing
macrophage activation. Tissue culture plastic (TCP) was also used as a control to which all data were compared using a two-way ANOVA. Hydrogen peroxide production from macrophages grown on any pSi– PCL composites was not significantly different to those grown on PCL but significantly lower than those on TCP, PMA or copper, whilst increasing pSi content in the composite proportionally decreased the macrophage response in terms of H2O2 production (c). (* P \ 0.05)
Macrophage studies clearly demonstrated that the pSi–PCL composites do not elicit a detrimental immune response. Morphological evidence of macrophage activation such as increased numbers of cell processes and surface membrane ruffling [35] was absent and biochemical measurements of hydrogen peroxide and IL-1b further demonstrated the lack of response. In fact, no material caused macrophages growing on the surface to secrete IL-1b (an important inflammatory cytokine) though the macrophages were able to produce IL-1b in dose dependant quantities when stimulated with LPS. Hydrogen peroxide
production remained low for macrophages cultured on tissue culture plastic, PCL and pSi–PCL composites and was significantly increased over 6 h for macrophages cultured on copper and PMA, surfaces known to elicit a macrophage response [36]. Macrophages actually produced less hydrogen peroxide on pSi–PCL composites than on tissue culture plastic, in a trend suggesting that increasing pSi content correlates with decreasing hydrogen peroxide production. Culturing primary human osteoblasts on the surface of polycaprolactone, pSi-composite discs and Thermanox
123
J Mater Sci: Mater Med
A
B
C
Fig. 5 The activity of human osteoblasts on the surface of polycaprolactone (PCL), pSi–PCL composites and thermanox coverslips (tissue culture plastic) was compared. a After 28 days, the DNA content of lysates from cells cultured on PCL-based materials was lower than Thermanox controls, but proportional to the amount of silicon in the composite (gradient:182 ng DNA/mg pSi). b The production of collagen by osteoblasts was again highest on
Thermanox, and lowest on the polycaprolactone-only disk; increasing pSi content in the composite resulted in an increased rate of collagen production. c After 28 days, the amount of collagen produced on composites containing 3 or 4.5 % pSi was significantly higher than polycaprolactone-only and equivalent to the amount produced on Thermanox. Error bars show standard error of the mean, n = 6 (composites), n = 12 (Thermanox), (*** P \ 0.001)
revealed that whilst Thermanox supports rapid cell growth in monolayer, the osteoblastic phenotype is less pronounced in terms of alkaline phosphatase expression and collagen production (when normalised for cell number). Conversely, polycaprolactone alone is a poor substrate for cell growth, but the addition of increasing quantities of pSi to the material resulted in proportional increases in cell proliferation, as seen with some other silicon-containing biomaterials [37]. Alkaline phosphatase expression steadily decreased over 28 days when cells were cultured on Thermanox, but exhibited peak activity at 14 days on the composites, with a subsequent decrease to a lower level of constitutive expression. This general trend in expression has been observed and reported by many biomaterials research groups, particularly those using primary osteoblasts [38, 39] and silicate/bioglass-based materials [40] and is also found in vivo [41]. Osteoblast production of collagen was also increased with the addition of pSi and, when normalised for cell number, collagen production was significantly higher on the pSi composite than Thermanox, indicating that the mechanism of silicon-induced osteoblast enhancement of osteogenesis
is via increased matrix production and cellular proliferation [37, 40, 41].
5 Conclusion In this work, the bioactivity of pSi—polycaprolactone composites has been characterised by a variety of means. Incorporating pSi with the polymer enhanced the deposition of hydroxyapatite on the surface of the composite over polycaprolactone controls, and this enhancement was found to be proportional to increasing silicon content. Furthermore, silicon-substituted hydroxyapatites appear to spontaneously form on the composites in simulated body fluid. Adding up to 8 per cent pSi to polycaprolactone increased the elastic modulus of the material and did not have any significant negative impact on the mechanical properties. At these concentrations, pSi supported osteoblast growth and appeared to reduce the response of cultured macrophages in vitro. For this initial investigation into polymer composites containing pSi, the biomaterial has been demonstrated to
123
J Mater Sci: Mater Med
possess similar characteristics to bioactive glass: forming silicic acids in solution which then polycondense to form surface layers of silica and apatite. It is the novel features of pSi which make it most interesting to the biomaterials community: having an open porous structure which can be surface modified, filled with pharmaceuticals and possessing optoelectronic properties derived from quantum phenomena. As silicon technology is so ubiquitous within the modern world, the development of pSi as a biomaterial may lead to fascinating interdisciplinary applications and provide further insights into the role of silicon in the biology of bone.
References 1. Carlisle EM. Silicon as an essential element. Fed Proc. 1974;33:1758–66. 2. Carlisle EM. Silicon as a trace nutrient. Sci Total Environ. 1988;73:95–106. 3. Reffitt DM, Ogston N, Jugdaohsingh R, Cheung HFJ, Evans BAJ, Thompson RPH, Powell JJ, Hampson GN. Orthosilicic acid stimulates collagen type 1 synthesis and osteoblastic differentiation in human osteoblast-like cells in vitro. Bone. 2003;32: 127–35. 4. Arumugam MQ, Ireland DC, Brooks RA, Rushton N, Bonfield W. Orthosilicic acid increases collagen type I mRNA expression in human bone-derived osteoblasts in vitro. Key Eng Mater. 2006;254:869–72. 5. Anderson SI, Downes S, Perry CC, Caballero AM. Evaluation of the osteoblast response to a silica gel in vitro. J Mater Sci Mater Med. 1998;9:731–5. 6. Xynos IE, Alasdair JE, Buttery LDK, Hench LL, Polak JM. Gene-expression profiling of human osteoblasts following treatment with the ionic products of Bioglass 45S5. J Biomed Mater Res. 2001;55:151–7. 7. Valerio P, Pereira MM, Goes AM, Leite MF. The effect of ionic products from bioactive glass dissolution on osteoblast proliferation and collagen production. Biomaterials. 2004;25:2941–8. 8. Sripanyakorn S, Jugdaohsingh R, Thompson RPH, Powell JJ. Dietary silicon and bone health. Nutr Bull. 2005;30:222–30. 9. Jugdaohsingh R, Tucker KL, Qiao N, Cupples LA, Kiel DP, Powell JJ. Dietary silicon intake is positively associated with bone mineral density in men and premenopausal women of the Framingham Offspring cohort. J Bone Miner Res. 2004;19: 297–307. 10. Hench LL. The story of BioglassÒ. J Mater Sci. 2006;17:967–78. 11. Canham T. Bioactive silicon fabrication via nanoetching techniques. Adv Mater. 1995;7:1033–7. 12. Anglin EJ, Cheng L, Freeman WR, Sailor MJ. Porous silicon in drug delivery devices and materials. Adv Drug Deliv Rev. 2008;60:1266–77. 13. Serda RE, Gu J, Bhavane RC, Liu X, Chiappini C, Decuzzi P, Ferrari M. The association of silicon microparticles with endothelial cells in drug delivery to the vasculature. Biomaterials. 2009;30:2440–8. 14. Zhang K, Loong SL, Connor S, Yu SW, Tan SY, Ng RT, Lee KM, Canham L, Chow PK. Complete tumor response following intratumoral 32P BioSilicon on human hepatocellular and pancreatic carcinoma xenografts in nude mice. Clin Cancer Res. 2005;11:7532–7.
123
15. Park JH, Gui L, Malzahn G, Ruoslahti E, Bhatia SN, Sailor MJ. Biodegradable porous silicon nanoparticles for in vivo applications. Nat Mater. 2009;8:331–6. 16. Coffer JL, Montchamp JL, Aimone JB, Weis RP. Routes to calcifled porous silicon: implications for drug delivery and biosensing. Physica Status Solidi A. 2003;197:336–9. 17. Anglin EJ, Schwartz MP, Ng VP, Perelman LA, Sailor MJ. Engineering the chemistry and nanostructure of porous silicon Fabry–Pe´rot films for loading and release of a steroid. Langmuir. 2004;20:11264–9. 18. Charnay C, Begu S, Tourne-Peteilh C, Nicole L, Lerner DA, Devoisselle JM. Inclusion of ibuprofen in mesoporous templated silica: drug loading and release property. Eur J Pharm Biopharm. 2004;57:533–40. 19. Vaccari L, Canton D, Zaffaroni N, Villa R, Tormen M, di Fabrizio E. Porous silicon as drug carrier for controlled delivery of doxorubicin anticancer agent. Microelectron Eng. 2006;83: 1598–601. 20. Salonen J, Kaukonen AM, Hirvonen J, Lehto V-P. Mesoporous silicon in drug delivery applications. J Pharm Sci. 2008;97:632–53. 21. Whitehead MA, Fan D, Mukherjee P, Akkaraju GR, Canham LT, Coffer JL. High-porosity poly(epsilon-caprolactone)/mesoporous silicon scaffolds: calcium phosphate deposition and biological response to bone precursor cells. Tissue Eng Part A. 2008; 14:195–206. 22. Johansson F, Kanje M, Linsmeier CE, Wallman L. The influence of porous silicon on axonal outgrowth in vitro. IEEE Trans Biomed Eng. 2008;55:1447–9. 23. Alvarez SD, Derfus AM, Schwartz MP, Bhatia SN, Sailor MJ. The compatibility of hepatocytes with chemically modified porous silicon with reference to in vitro biosensors. Biomaterials. 2008;30:26–34. 24. Dash TK, Konkimalla VB. Poly-E-caprolactone based formulations for drug delivery and tissue engineering: a review. J Control Release. 2012;158:15–33. 25. Sinha VR, Trehan A. Development, characterization, and evaluation of ketorolac tromethamine-loaded biodegradable microspheres as a depot system for parenteral delivery. Drug Deliv. 2008;15:365–72. 26. Rohner D, Hutmacher DW, Cheng TK, Oberholzer M, Hammer B. In vivo efficacy of bone-marrow-coated polycaprolactone scaffolds for the reconstruction of orbital defects in the pig. J Biomed Mater Res B. 2003;66:574–80. 27. Halimaoui A. Porous silicon formation by anodization. In: Canham LT, editor. Properties of porous silicon. London: Institution of Engineering and Technology; 1997. 28. Kokubo T, Kushitani H, Sakka S, Kitsugi T, Yamamuro T. Solutions able to reproduce in vivo surface-structure changes in bioactive glass-ceramic A-W. Biomed Mater Res. 1990;24:721–34. 29. Chouzouri G, Xanthos M. In vitro bioactivity and degradation of polycaprolactone composites containing silicate fillers. Acta Biomater. 2007;3:745–56. 30. Venugopal JR, Low S, Choon AT, Kumar AB, Ramakrishna S. Nanobioengineered electrospun composite nanofibers and osteoblasts for bone regeneration. Artif Organs. 2008;32:388–97. 31. Porter AE, Patel N, Skepper JN, Best SM, Bonfield W. Comparison of in vivo dissolution processes in hydroxyapatite and silicon-substituted hydroxyapatite bioceramics. Biomaterials. 2003;24:4609–20. 32. Porter AE, Patel N, Skepper JN, Best SM, Bonfield W. Effect of sintered silicate-substituted hydroxyapatite on remodelling processes at the bone-implant interface. Biomaterials. 2004;25: 3303–14. 33. Gibson IR, Best SM, Bonfield W. Effect of silicon substitution on the sintering and microstructure of hydroxyapatite. J Am Ceram Soc. 2002;85:2771–7.
J Mater Sci: Mater Med 34. Schwarz K. A bound form of silicon in glycosaminoglycans and polyuronides. Proc Natl Acad Sci. 1973;70: 1608–12. 35. Kleveta G, Borze˛cka K, Zdioruk M, Czerkies M, Kuberczyk H, Sybirna N, et al. LPS induces phosphorylation of actin-regulatory proteins leading to actin reassembly and macrophage motility. J Cell Biochem. 2012;113:80–92. 36. Scotchford CA, Garle MJ, Batchelor J, Bradley J, Grant DM. Use of a novel carbon fibre composite material for the femoral stem component of a THR system: in vitro biological assessment. Biomaterials. 2003;24:4871–9. 37. Keeting PE, Oursler MJ, Weigand KE, Bonde SK, Spelsberg TC, Riggs BL. Zeolite A increases proliferation, differentiation and transforming growth factor beta production in normal adult human osteoblast-like cells in vitro. J Bone Miner Res. 1992;7: 1281–9.
38. Sammons RL, El Haj AJ, Marquis PM. Novel culture procedure permitting the synthesis of proteins by rat calvarial cells cultured on hydroxyapatite particles to be quantified. Biomaterials. 1994;15:536–42. 39. Homaeigohar SSh, Shokrgozar MA, Sadi AY, Khavandi A, Javadpour J, Hosseinalipour M. In vitro evaluation of biocompatibility of beta-tricalcium phosphate-reinforced high-density polyethylene; an orthopedic composite. J Biomed Mater Res A. 2005;75:14–22. 40. Kubo K, Tsukasa N, Uehara M, Izumi Y, Ogino M, Kitano M, Sueda T. Calcium and silicon from bioactive glass concerned with formation of nodules in periodontal-ligament fibroblasts in vitro. J Oral Rehabil. 1997;24:70–5. 41. Zhou H, Choong P, McCarthy R, Chou ST, Martin TJ, Ng KW. In situ hybridization to show sequential expression of osteoblast gene markers during bone formation in vivo. J Bone Miner Res. 1994;9:1489–99.
123
