Highly Porous Amorphous Calcium Phosphate for Drug Delivery and Bio-Medical Applications

Amorphous calcium phosphate (ACP) has shown significant effects on the biomineralization and promising applications in bio-medicine. However, the limited stability and porosity of ACP material restrict its practical applications. A storage stable highly porous ACP with Brunauer–Emmett–Teller surface area of over 400 m2/g was synthesized by introducing phosphoric acid to a methanol suspension containing amorphous calcium carbonate nanoparticles. Electron microscopy revealed that the porous ACP was constructed with aggregated ACP nanoparticles with dimensions of several nanometers. Large angle X-ray scattering revealed a short-range atomic order of <20 Å in the ACP nanoparticles. The synthesized ACP demonstrated long-term stability and did not crystallize even after storage for over 14 months in air. The stability of the ACP in water and an α-MEM cell culture medium were also examined. The stability of ACP could be tuned by adjusting its chemical composition. The ACP synthesized in this work was cytocompatible and acted as drug carriers for the bisphosphonate drug alendronate (AL) in vitro. AL-loaded ACP released ~25% of the loaded AL in the first 22 days. These properties make ACP a promising candidate material for potential application in biomedical fields such as drug delivery and bone healing.


S1.1. Large angle X-ray scattering (LAXS)
The LAXS measurements of ACP samples used the same method for HPACC as detailed in our previous work [1].
The reduced intensity function is Fourier transformed in order to get the radial distribution function (RDF) according to the following equation:
For the exposure experiments, cells were seeded in clear 96-well tissue plates at a density of 3200 cells per well (200 µl) and incubated for 24 hours. The cell culture medium in the wells was removed after the initial incubation time and material dispersions of ACP032 and ACP053, at concentrations of 500 µg/mL, 200 µg/mL, 100 µg/mL, 50 µg/mL and 25 µg/mL, in cell culture medium were added to the wells. Untreated cells were the negative control and cells exposed to cell culture medium supplemented with 5 v/v% DMSO served as the positive control. Cells were further cultured for 24 ± 2 hours and 48 ± 2 hours at 37˚C, in 5% CO2 and a humidified atmosphere. The cell culture medium was removed from the wells after the incubation period and the cells were washed with PBS. 200 µL of a 10 v/v% presto blue solution in cell culture medium was added to each well and incubated for 60 min. 100 µL from each well was then transferred to black 96-well plates and the fluorescence was measured at wavelengths of λex. = 560 nm and λem. = 590 nm using a Tecan Infinite M200 spectrofluorometer (Männedorf, Switzerland). MC3T3 cells exposed to ACP032 and ACP053 dispersions at 25 µg/mL and 500 µg/mL of each respective material, untreated cells and cells exposed to 5% DMSO were stained with calceinacetoxymethyl (AM) and propidium iodide (PI) and thereafter visualized by fluorescence microscopy. Briefly, after exposure for 24 ± 2 hours and 48 ± 2 hours, the cell culture medium in the wells was removed and 100 µL of a dye solution consisting of 2 µL calcein-AM and 1 µL PI per ml of cell culture medium was added per well. After incubation at 37˚C in 5% CO2 and a humidified atmosphere for 15 min, the stained cells were imaged using a Nikon Eclipse TE2000 microscope (Minato, Tokyo, Japan) with a filter set at λex. = 490 nm, λem. = 515 nm for viable cells and λex. = 535 nm, λem. = 617 nm for dead cells.
Cell viability data were log-transformed and analyzed using Welch's ANOVA followed by the Games-Howell post-hoc test using R Studio software (v.3.5.2), where differences from the negative control (untreated cells) were considered statistically significant when p was < 0.05. All experiments (n = 6) were performed in triplicate. S1.3. Drug load and in vitro release AL was loaded into ACP by the soaking method. Specifically, 1.0 g ACP was added to 150 mL methanol solution containing AL at a concentration of 90 µg/mL. After shaking at 200 rpm on a shaker (BioSan Orbital Shaker PSU-10i, Nordic Biolabs) for 1 day, the dispersion was centrifuged at 1357 × g for 30 min. The ACP was recovered and the AL loading procedure was repeated five times. The concentration of AL in the supernatant after each loading step was recorded by colorimetry based on the reaction of the primary amino group in AL with ninhydrin in a methanol medium in the presence of 0.05 M sodium bicarbonate [2]. The color product had UV absorption at 568 nm on a UV-vis spectrophotometer (Shimadzu UV-1800 spectrophotometer with 1 cm matched cells). After carrying out the loading procedure five times, the obtained ACP-AL was dried at 70 °C in a ventilated oven for 24 hours.
The calibration curve was made from four solutions with concentration 6.2542 µg/mL, 8.3390 µg/mL, 12.5085 µg/mL and 25.0170 µg/mL. These four concentrations were in the linearity range (3.75-45 µg/mL) in the reference [2]. The acquired calibration curve was y = 0.02396x + 0.00203, (y is the instrument response, i.e. absorbance and x is the concentration in unit of µg/mL, R2=0.99999). The limit of detection (LOD) was further calculated by LOD = 3.3 x σ / S (S is the slope of the calibration curve and σ is the standard deviation of the response) [3]. The LOD of the method used in this study was 0.058 µg/mL. The loading content of AL in ACP was determined by the difference in the concentration of the AL-methanol solution before and after loading process. The original concentration of AL-methanol concentration is 90 µg/mL. The concentration of the solution after loading was tested by colorimetry method as discussed previously. The drug loading efficiency, as the percentage of AL mass loaded into ACP with respect to the initial AL mass was also calculated.
The in vitro dissolution tests were carried out in HEPES aqueous buffer (10 mM, pH = 7.0-7.6) (AL has solubility in water range from 10 mg/mL to 50 mg/mL [4][5][6]). 210 mg of AL-loaded ACP (11.76 mg AL) was added to 120 mL of 10 mM HEPES buffer in a glass laboratory bottle. The mixture was kept shaking at 200 rpm on a shaker (BioSan Orbital Shaker PSU-10i, Nordic Biolabs) at 37 °C in an oven. Aliquots of 1.2 mL were withdrawn from each vessel at regular time intervals and centrifuged at 14000 rpm for 10 min, and then 1 mL of the aliquot was used to test the amount of AL released by the colorimetric method detailed above [2]. The remaining 0.2 mL was added to 1.0 mL of fresh HEPES buffer and was reintroduced back into the bottle. The drug release experiment was repeated three times, and all measurements were performed in triplicate. The average concentration and the corresponding standard deviations were calculated.
In the thermogravimetric analysis curves for CaP068, CaP078 and CaP088 ( Figure S2a), the mass loss between 300 and 500 °C was attributed to the condensation reaction of monetite [7].

2CaHPO4  Ca2P2O7 + H2O
The magnitude of the mass loss was in the order CaP088 > CaP078 > CaP068. This agreed well with the content of monetite inside the samples. CaP078 and CaP088 were in the crystalline form (monetite), as seen from the powder PXRD patterns in Figure 1, and therefore there were no exothermic peaks in the DSC curves, only endothermal peaks between 300 and 410 °C representing the condensation of monetite. Note that there was no indication of mass loss related to the decomposition of calcium carbonate in the TGA curves of CaP068, CaP078 and CaP088. The carbonate content in ACP032 was calculated as detailed below: Weight loss up to 300 °C was assumed to be related to the removal of adsorbed or structural water within ACP032. The weight loss between 600 and 800 °C was attributed to the mass of CO2 released from the decomposition of CaCO3 to CaO. This weight loss can be used to back calculate the amount of CaCO3 presence in ACP032. The carbonate content was determined to be ~13.0 wt.% of the starting material (including adsorbed water) and ~15.0 wt.% of the dried ACP032. The calculated carbonate content using this method was close to the value obtained from ICP-OES.  Figure S4 shows the XPS spectra for ACP053 and ACP032. Ca, O, P and C were detected in both samples. The Ca/P atomic ratios in ACP053 and ACP032 were 1.23 and 2.23, respectively, which was in agreement with the ICP-OES results. Three different kinds of carbon were detected in the highenergy resolution XPS C1 s spectrum for ACP032: C-C/C=C at 284.8 eV, adventitious carbon at 287.1 eV and carbon in CO32-at 289.2 eV. However, in ACP032, only C-C at 284.8 eV and carbon in CO32at 289.9 eV were detected. It should be noted that the relative intensity of carbon in CO32-was much lower in ACP053 than in ACP032 because of the low CO32-content in ACP053. The high-energy resolution XPS O 1s spectrum for ACP053 suggested that O existed in PO43-at 531.2 eV and H2O at 533.5 eV while, in the spectrum for ACP032, O was in PO43-at 531.2 eV, H2O at 533.1 eV and carbonate at 532.1 eV. The Ca 2p spectrum for ACP053 could be divided into Ca 2p3/2 at 347.3 eV and Ca 2p1/2 at 350.9 eV, similar to the Ca 2p spectrum for ACP032 (Ca 2p3/2 at 347.4 eV and Ca 2p1/2 at 351.0 eV). The P 2p spectrum for ACP053 was located at 133.2 eV with P 2p3/2 at 132.9 eV and P 2p1/2 at 133.9 eV while that for ACP032 was at 133.4 eV with P 2p3/2 at 133.0 eV and P 2p1/2 at 133.9 eV.     Figure S7d is the corresponding selectedarea electron diffraction pattern).    S3.2. Stability in de-ionized water Figure S12. (a) Powder X-ray diffraction pattern for the amorphous calcium carbonate sample ACP053 after exposure to de-ionized water for 15 hours (donated as ACP053-W-15H), also showing the diffraction peaks for hydroxyapatite (PDF 00-009-0432) and (b) infrared spectra for ACP053 after immersion in de-ionized water for 1-15 hours. Figure S13. Scanning electron microscopy images of the hydroxyapatite formed from amorphous calcium phosphate sample ACP053 after exposure to de-ionized water for 15 hours.
The stability of ACP032 after exposure to de-ionized water was explored using the same methods as for ACP053. Infrared (IR) spectra for ACP032 after exposure to de-ionized water for 1 hour to 50 days are shown in Figure 14a. The spectrum for ACP032 after exposure to water for 20 hours had the ν2 band for CO32-at around 866.6 cm-1 and the ν4 band for PO43-at 582.5 cm-1 and 560.6 cm-1, similar to those for as-synthesized ACP032, indicating that it had remained amorphous for over 20 hours. After exposure to de-ionized water for 30 hours, the ν2 band for CO32-became sharp and shifted to ~870 cm-1, which suggested that ACP032 had started to crystallize. The PXRD pattern confirmed that calcite had formed at this stage ( Figure S15). The ν4 band for PO43-began to split into individual bands after ACP032 had been immersed in de-ionized water for 15 days. The bands became well-defined after 30 days. These bands were attributed to HA, which suggested that ACP032 had crystallized to HA after exposure to de-ionized water for 15 days and that this crystallization process proceeded over 30 days. The IR spectra for ACP032 after exposure to de-ionized water for 1 hour and 30 days, and for calcite and HA are shown in Figure S14b. The ν4 and ν2 bands for CO32-and the ν4 band for PO43-were used to follow the crystallization process of ACP032. The PXRD pattern for ACP032 after exposure to de-ionized water for 30 days demonstrated that the formed crystalline phases were calcite (PDF 00-002-0623) and HA (PDF 00-001-1008 and PDF 00-066-0147), as shown in Figure S15b.  Figure S14. (a) Infrared (IR) spectra for the amorphous calcium phosphate sample ACP032 after exposure to de-ionized water for 1 hour to 50 days and (b) IR spectra for ACP032 after exposure to deionized water for 1 hour and 30 days, in comparison to those for hydroxyapatite and calcite.  Figure S15. Powder X-ray diffraction patterns for the amorphous calcium phosphate sample ACP032 after exposure to de-ionized water for (a) 1 hour to 50 days and (b) 30 days (ACP032-W-30D), showing the diffraction peaks for calcite (PDF 00-002-0623), hydroxyapatite (PDF 00-001-1008) and carbonated hydroxyapatite (PDF 00-066-0147).   Figure S20. (a) N2 sorption isotherms for the amorphous calcium phosphate sample ACP053 before and after loading with alendronate (AL) and (b) the corresponding density functional theory pore-size distribution graphs. Pore volume from N2 adsorption (cm 3 /g) -1.11