Shape-dependent toxicity and mineralization of hydroxyapatite nanoparticles in A7R5 aortic smooth muscle cells

Vascular smooth muscle cell damage is a key step in inducing vascular calcification that yields hydroxyapatite (HAP) as a major product. The effect of the shape of HAP on the damage to vascular smooth muscle cells has yet to be investigated. In this study, we compared the differences in toxicity of four various morphological nano-HAP crystals, namely, H-Rod, H-Needle, H-Sphere, and H-Plate, in rat aortic smooth muscle cells (A7R5). The sizes of these crystals were 39 nm × 115 nm, 41 nm ×189 nm, 56 nm × 56 nm, and 91 nm × 192 nm, respectively. Results showed that all HAPs decreased cell viability, disorganized cell morphology, disrupted cell membranes, increased intracellular reactive oxygen species concentration, decreased mitochondrial membrane potential, decreased lysosome integrity, increased alkaline phosphatase activity, and increased intracellular calcium concentration, resulting in cell necrosis. The cytotoxicity of the four kinds of HAP was ranked as follows: H-Plate > H-Sphere > H-Needle > H-Rod. The cytotoxicity of each crystal was positively correlated with the following factors: large specific surface area, high electrical conductivity and low surface charge. HAP accelerated calcium deposits on the A7R5 cell surface and induced the expression of osteogenic proteins, such as BMP-2, Runx2, OCN, and ALP. The crystals with high cytotoxicity caused more calcium deposits on the cell surface, higher expression levels of osteogenic protein, and stronger osteogenic transformation abilities. These findings elucidated the relationship between crystal shape and cytotoxicity and provided theoretical references for decreasing the risks of vascular calcification.

Intracellular reactive oxygen species (ROS) detection assay. Following the cell incubations described as above, the cells were exposed to 200 μg/mL HAP crystals with various shapes for 24 h, the samples were stained with DCFH-DA for 20 min. The cells were then observed under a fluorescence microscope (Leica DMRA2, Germany). The fluorescence intensity was detected by using a microplate reader 20 .
Observation and detection of lysosomal integrity. Following the cell incubations described as above, the cells were loaded with 5 µg/mL acridine orange (AO) in DMEM for 15 min and incubated with 200 μg/mL HAP crystals with various shapes for 24 h. The distribution of AO was observed under a fluorescence microscope. The fluorescence intensity was measured with excitation at 485 nm and emission at 530 nm (green cytoplasmic AO) and 620 nm (red lysosomal AO) 16 . Measurement of mitochondrial membrane potential (Δψm). Following the cell incubations described as above, the cells were incubated with 200 μg/mL HAP crystals with various shapes for 24 h. The samples were stained with JC-1 stain, and 2 × 10 4 cells were detected through flow cytometry (FACS Aria, BD Corporation, CA, USA).
Cell apoptosis and necrosis detection. Following the cell incubations described as above, the cells were incubated with 200 μg/mL HAP crystals with various shapes for 24 h. The samples were stained with 5 µL of Annexin V-FITC for 10 min in dark, and then stained with 5 µL propidium iodide. Finally, 2 × 10 4 cells were detected through flow cytometry.
Observation of calcified nodules through alizarin red staining. The medium of each group was replaced every 2 days and incubated with DMEM containing 1% FBS for 14 days. After the treatments were administered, the cells were fixed with paraformaldehyde for 20 min, incubated with 0.1% alizarin red staining (pH = 4.2) for 0.5 h, washed the cells, and observed under a microscope (magnification, 100×). HAP, when used, was applied from days 0 to 1. Then the extracellular HAPs of the cells were removed and the medium containing 1% FBS were used to culture the cells.
Quantitative analysis: The cells were fixed with 70% ethanol for 1 h, washed the cells, and stained with 0.1% alizarin red solution (pH = 4.2) for 1 h. Then the cells were incubated with PBS for 15 min, washed thrice with PBS, and incubated in 10% (w/v) cetylpyridinium chloride for 30 min. Absorbance was detected through a microplate reader at 562 nm, and the absorbance of the supernatant of a group of simple cells without alizarin red was determined.
Intracellular and extracellular distribution of HAP. After the cells were exposed to 200 μg/mL FITC-HAP crystals with various shapes for 6 h, this time was required to complete endocytosis without causing too much damage www.nature.com/scientificreports www.nature.com/scientificreports/ to the cells. The cell membrane was stained with 300 μL of 10 μM DiI for 15 min, and the cells were fixed with paraformaldehyde. Then DAPI staining solution was added to stain the cells for 5 min. The prepared samples were observed under a confocal microscope (LSM510 Meta Duo Scan, Zeiss, Germany).
Quantitative analysis of internalized HAP crystals. After the cells were exposed to 200 μg/mL FITC-HAP crystals with various shapes for 6 h, the supernatant was removed. The cells were treated with 0.4 mL of EDTA (5 mM) for 5 min to remove the bound HAP 21 . Then, the cells were rinsed thrice with cold PBS to completely remove the external soluble HAP, and then detected through flow cytometry. The number of cells analyzed in the flow cytometry experiments was 2 × 10 4 .
Detection of intracellular calcium concentration. After 24 h of incubation, the cells were stained with 200 µL of Fluo-4/AM staining, incubated at 37 °C for 30 min, washed thrice with PBS, and detected through flow cytometry. The number of cells analyzed in the flow cytometry experiments was 2 × 10 4 .
ALP activity assay. The medium of each group was replaced every 2 days and incubated with 1% DMEM containing 1% FBS for 14 days. After the treatments were administered, the cells were fixed with paraformaldehyde for 20 min and incubated with ALP staining in accordance with the manufacturer's instructions. The stained cells were observed by phase contrast microscope (magnification, 200×). Blue stain indicated a high ALP activity.
Quantitative analysis: After 14 days of incubation, the ALP activity was assessed in the supernatants by using an ALP assay kit. Protein concentrations were detected by a bicinchoninic acid (BCA) protein assay kit, and the ALP activity was normalized for cellular protein content.
Osteogenic protein expression by western blotting assays. Cell lysate was prepared through lysis buffer. Equal amounts of protein were loaded and separated on 12% SDS-PAGE and transferred to a PVDF membrane. The membranes were then incubated with primary antibodies against human BMP-2, Runx2, and OCN overnight and detected with the secondary HRP-conjugated antibody. Immune complexes were visualized using an ECL system. Western blot signal intensities were quantified using AlphaEaseFC (Alpha Innotech, San Leandro, CA, USA). The integrated density values for the test and control bands were obtained and shown as their ratio.
Statistical analysis. Experimental data were expressed as mean ± standard deviation (x ± SD). Experimental results were analyzed statistically using SPSS 13.0 (SPSS Inc., Chicago, IL, USA). Differences in the means between the experimental groups and the control group were analyzed using one-way ANOVA, followed by www.nature.com/scientificreports www.nature.com/scientificreports/ Tukey's post hoc test. p < 0.05 indicated significant differences, p < 0.01 corresponded to extremely significant difference, and p > 0.05 denoted no significant differences.    Figure 1B illustrates the FT-IR of HAP crystals with various shapes. A broad absorption peak near 3424 cm −1 belonged to water adsorbed on the surface of the HAP nanoparticles; the absorption peak near 3575 cm −1 belonged to the O-H stretching vibration in HAP; and the vibration peaks at 567, 605, 959, and 1036 cm −1 were attributed to the asymmetric stretching vibration peaks of P-O in the PO 4 3− groups 22 . XRD and FT-IR spectroscopy results showed that all of the four HAP nanoparticles were in their pure phase. Figure 1C shows the SEM images of the four HAP crystals, namely, rod-like HAP (H-Rod), needle-like HAP (H-Needle), sphere-like HAP (H-Sphere), and plate-like HAP (H-Plate).

Results
Zeta potential and conductivity detection of HAP. The zeta potential values of HAP dispersed in pure water and culture medium were negative ( Table 1). The absolute value of the zeta potential in the medium obviously reduced because a large number of inorganic ions (e.g., Ca 2+ and Mg 2+ ), amino acids, vitamins, and other auxiliary components are present in DMEM [23][24][25] . Therefore, the exposed PO 4 3− on the surface of the HAP crystal strongly interacted with Ca 2+ and Mg 2+ in the medium, resulting in the partial neutralization of PO 4 3− . When PO 4 3− adsorbed a high concentration of cations, the absolute value of the zeta potential decreased.
HAP is a poorly soluble substance, and it is slightly soluble in pure water. The conductivity of HAP in pure water is low (0.22-0.75 mS/cm), whereas the conductivity of the four HAP nanoparticles in the medium is obviously higher (14.0-16.1 mS/cm) than that in pure water possibly because of the large amount of inorganic ions (e.g., Ca 2+ and Mg 2+ ) and amino acids in the medium 23 .
S BET , pore volume, and pore size of HAP. The adsorption and desorption curves of HAP nanoparticles with various shapes are shown in Fig. 2. The curve of H-Rod is a typical Ι-type (micropore) adsorption isotherm 26 , and its S BET and pore size are small (25.04 m 2 /g and 3.14 nm; Table 1). The adsorption and desorption curves of other crystals were a typical type III adsorption isotherm, and the curves tended to be directed to the X axis in the low-pressure section, indicating that the interactions between N 2 and these crystals were weak, and their surface with holes were rough and had pore sizes of 30.45, 35.29, and 13.44 nm for H-Needle, H-Sphere,  Figure 3A shows the changes in the viability of A7R5 cells treated with four types of HAP for 24 h. The four types of HAP elicited an obvious toxic effect on A7R5 cells in a concentration-dependent manner. The order of cytotoxicity was as follows: H-Plate > H-Spher e > H-Needle > H-Rod. At the crystal concentration of 200 μg/mL, the cell viabilities of H-Plate-, H-Sphere-, H-Needle-, and H-Rod-treated groups were 56.88%, 65.14%, 70.31%, and 76.03%, respectively, which were significantly lower than those of the control group (p < 0.01). The results of the analysis of the correlation of cytotoxicity with S BET and electrical conductivity are shown in Fig. 3C,D, respectively. Their correlation coefficients were 0.8384 and 0.8425, respectively. The order of influence of different physical and chemical properties of HAP on its cytotoxicity was as follows: electrical conductivity > S BET > Zeta potential. cell membrane damage induced by HAp with various shapes. LDH is a stably present cytosolic enzyme. When the cell membrane is damaged, LDH is released outside the cell, and the increased release of intracellular LDH is considered an important indicator of cell membrane integrity.

Toxicity of HAP with various shapes on A7R5 cells.
In Fig. 3B, all of the four kinds of HAP caused an increase in intracellular LDH release in different degrees and showed a concentration-dependent manner. LDH release was ranked in the following order: H-Plate > H-Spher e > H-Needle > H-Rod. This rule was consistent with cell viability (Fig. 3A). Figure 4 shows the morphological changes in A7R5 cells treated with four different morphological HAP nanoparticles. In the control group, the cell showed a www.nature.com/scientificreports www.nature.com/scientificreports/ plump spindle shape, the morphologies of the cells treated with HAP crystals became disordered. The cell density was obviously reduced. The tight junctions between the cells were destroyed, and some cells had an expanded cytoplasm. The cell damage induced by H-Plate was the most serious. The orders of crystal damage to cell morphology were as follows: H-Plate > H-Sphere > H-Needle > H-Rod. Fig. 5, the HAP treatment increased ROS generation in A7R5 cells. The cells in the normal group had almost no green fluorescence (Fig. 5A), indicating that the intracellular ROS level was low (125). The green fluorescence of H-Rod-(266), H-Needle-(359), H-Sphere-(463), and H-Plate (519)-treated cells gradually increased, indicating that their intracellular ROS levels gradually increased (Fig. 5B). High cellular ROS levels may lead to apoptosis or necrosis 27 . changes in the lysosome integrity in the cells treated with HAp with various shapes. The degree of lysosomal damage can be determined by Acridine orange (AO) dye 28 . As shown in Fig. 6, the cells in the normal group retained an intact lysosome structure. The emitted red fluorescence of the lysosomes merged with the green fluorescence of the cytoplasm, thereby presenting orange fluorescence. When normal cells were damaged by HAP nanoparticles, the red fluorescence was obviously reduced. The damage to the lysosomes in A7R5 cells treated with the four types of HAP nanoparticles showed the following trend: H-Plate > H-Sphere > H-Needle > H-Rod.

Effects of HAP with various shapes on Δψm.
A decreased Δψm is a hallmark of early cell death.
The degree of Δψm can be determined by JC-1 dye 29 . In Fig. 7, the ratio of the cells with low ΔΨm (green fluorescent) to the normal cells was 3.12%, whereas the ratio of the cells in the HAP crystal-damaged group to the www.nature.com/scientificreports www.nature.com/scientificreports/ cells in the control group obviously increased (6.6%-19.48%), indicating that the different morphologies of HAP caused varying degrees of mitochondrial depolarization. The ratios of low-potential cells in the H-Rod, H-Needle, H-Sphere, and H-Plate groups were 6.6%, 8.38%, 12.96%, and 19.48%, respectively. Apoptosis or necrosis induced by HAp with various shapes. Cell apoptosis and necrosis were quantified through Annexin V/PI double staining 30 (Fig. 8). The cells treated with the four types of HAP exhibited varying degrees of cell necrosis rather than cell apoptosis compared with those of the control group (Fig. 8A,B). The necrosis rates (Q1 + Q2) of the cells treated with H-Plate, H-Sphere, H-Needle, and H-Rod for 24 h were 17.13%, 16.47%, 12.41%, and 9.67%, which respectively increased to 38.1%, 31.59%, 19.08%, and 15.97% when the treatment time was extended to 14 days. The degree of cell necrosis induced by H-Plate and H-Sphere was greater than that induced by H-Rod and H-Needle.

Calcium depositions on A7R5 cells treated with HAP with various shapes. Alizarin red chelates
with calcium to form orange-red calcium deposits 31 . Vascular smooth muscle cell injury is a key step in inducing vascular calcification 32 . In Fig. 9, no obvious calcium depositions in normal cells were observed, but the four kinds of HAP caused different degrees of calcium deposition. The changes in the calcium deposition contents were described as follows: H-Plate > H-Sphere > H-Needle > H-Rod.
HAP distribution inside and outside A7R5 cells. HAP crystals were labeled with FITC (green fluorescence) to verify HAP distribution inside and outside A7R5 cells 33 . The control cells exhibited a typical spindle shape with a full and intact morphology, whereas the HAP crystals caused morphological disorder of cells. A7R5 cells internalized all of the crystals that adhered to the membrane (Fig. 10). The green crystals appeared yellow after they adhered to the red cell membrane. A large yellow area corresponded to the presence of numerous crystals. HAP can encapsulate in cells by vesicles 34,35 . In our study, the internalization degree was higher in H-Sphere and H-Plate than in H-Rod and H-Needle. www.nature.com/scientificreports www.nature.com/scientificreports/ internalization degree of HAp nanoparticles with various shapes. After the A7R5 cells were exposed to the FITC-labeled HAP crystals with varying shapes for 6 h, the bound HAP was removed by using 0.4 mL of EDTA 19 . The proportion of the cells with endocytic crystals was examined through flow cytometry (Fig. 11A). The cells with positive FITC signaling can be considered as cells with endocytic crystals. All of the four types of HAP crystals could be endocytosed by A7R5 cells. The percentage of the cells with endocytic crystals was ranked in the following order: H-Plate > H-Sphere > H-Needle > H-Rod (Fig. 11B).
The cells with endocytic crystals for H-Plate was greater than that of H-Sphere probably because flat nanoparticles easily penetrate the lipid bilayer of the cell membrane and successfully enter cells; conversely, spherical nanoparticles tend to affect the membrane, prefer to stay near the membrane center, and stay on the membrane for a long time before they enter cells 36 . changes in intracellular ca 2+ in A7R5 cells treated with HAP with various shapes. Excessive generated Ca 2+ can cause cell damage and even cell death 37 . Fluo-4/AM is a fluorescent dye that can penetrate cell membranes. Fluo-4/AM that successfully enters a cell can be cleaved by intracellular esterase to form Fluo-4. Fluo-4 can combine with Ca 2+ to produce strong green fluorescence. Thus, the concentration of intracellular Ca 2+ can be determined by detecting the percentage of Furo-4 positive cells 38 .
In Fig. 12, the four types of HAP caused an increase in intracellular Ca 2+ compared with that in the normal group. The intracellular Ca 2+ fluorescence ratios in the control group, H-Rod, H-Needle, H-Sphere, and H-Plate were 3.14%, 6.09%, 10.82%, 13.14%, and 19.85%, respectively. Figure 12C shows the correlation analysis result. HAP endocytosis by the cells was positively correlated with intracellular Ca 2+ (R 2 = 0.9688). A high degree of crystal endocytosis corresponded to a high intracellular calcium concentration.

Changes in ALP activity in A7R5 cells treated with HAP with various shapes. ALP is an early
marker of osteoblast formation 10 . The ALP expression in normal VSMCs is low, and its expression levels in calcified blood vessels and heart valves are obviously high. The four HAP nanoparticles promoted the activity of ALP (Fig. 13)  www.nature.com/scientificreports www.nature.com/scientificreports/ High expression of BMP-2, Runx2 and OCN in A7R5 cells. BMP-2 is one of the most important extracellular signaling molecules that promote bone formation and induce osteoblast differentiation. As a target gene of BMP-2, Runx2 is also an important regulator of osteoblast differentiation and bone development, and OCN is regarded as an osteogenesis marker 10 .
NPS-2134 is a common Ca-sensing receptor inhibitor. NPS-2143 can block the expression of CaSR, which can further affect the expression of BMP-2 and Runx2. The protein expression levels of BMP-2, Runx2, and OCN were assayed via Western blot analysis. As shown in Fig. 14, after 14 days of treatment with different morphological HAP groups, the protein levels of BMP-2, Runx2, and OCN significantly increased, and the osteogenic transformation induced by HAP-Plate and HAP-Sphere in A7R5 cells was more severe than that caused by HAP-Rod and HAP-Needle (Fig. 14B-D). Figure 14 also shows that the differentiation of A7R5 cells to osteoblast-like cells under calcification conditions was mediated by CaSR-stimulated BMP-2 and Runx2 signaling pathways. Under the action of BMP-2 and Runx2 osteogenic factors, cells expressed excessive osteogenic proteins (OCN and ALP), induced osteogenic transformations, and increased the risks of vascular calcification.

Discussion
Toxicity mechanism of HAP with various shapes on A7R5 cells. Extracellular LDH levels are important indicators to verify changes in cell membrane permeability 39 . Adherent HAP crystals caused the release of LDH (Fig. 3B), demonstrating that HAP caused an increase in cell membrane permeability. Cell membrane rupture can cause intracellular electrolyte disorders, produce large amounts of ROS (Fig. 5). Excessive ROS generation by exogenous particles can induce oxidative stress 40 , which is a vital mechanism of cell toxicity 41 . Excessive ROS formation can also induce a decrease in Δψm (Fig. 7) 42 . Decreased Δψm often precedes cellular apoptosis and necrosis 43 .
HAP crystals were endocytosed by A7R5 cells (Figs. 10 and 11), leading to cell membrane rupture (Fig. 3B). Endocytosis occurs after HAP crystals interact with cells for 1 h 44 . Therefore, HAP may damage the cell membrane after crystal endocytosis. And cell membrane damage also caused by the adhered HAP 45 . Extracellular particles can enter cells through macrophagocytosis and membrane rupture 46 . www.nature.com/scientificreports www.nature.com/scientificreports/ Lysosomes contain many acidic hydrolyzing enzymes, and their pH is approximately 4.5 47 . In our study, the HAP crystals were dissolved via the acid hydrolyzing enzymes, thereby causing a remarkable increase in intracellular Ca 2+ (Fig. 12), which destroyed the osmotic pressure balance on lysosomal membranes, causing excessive lysosomal disruption and cell necrosis (Fig. 8) 48 .
The intracellular Ca 2+ production caused by HAP with various shapes differed. High amounts of H-Plate and H-Sphere released high Ca 2+ contents, resulting in high rates of cell necrosis (Fig. 8).
Potential vascular calcification risk differences caused by HAP with different shapes. The mechanism of cell injury and calcification caused by HAP nanoparticles with various shapes is summarized in Fig. 15. Cell necrosis may be more likely to cause vascular calcification than apoptosis 49 . The increasing cell volume caused cell to rupture and produce a large number of membrane-like necrotic fragments, which promote the formation of HAP crystals. Thereby, H-Plate and H-Sphere generated more calcium deposits (Fig. 9). The increase in intracellular Ca 2+ was also an important cause of the activation of HAP formation. Massive Ca 2+ caused the depletion of calcium inhibitor and the exposure of the protein annexin AnxA6/phosphatidylserine to the matrix vesicle surface, which can provide nucleation sites for HAP and increase calcium deposition 50 . The deposited HAP on the cell surface and the endocytosed crystals in the cells led to excessive ROS production [51][52][53] , thereby increasing the ALP activity and eventually mediating the osteogenic differentiation. HAP promotes the expression of osteogenic protein, and BMP-2 and Runx2 have synergistic effects on osteoblast differentiation. Runx2 can promote the expression of ALP and OCN 54 . Therefore, the production of membrane fragments caused by www.nature.com/scientificreports www.nature.com/scientificreports/ cell necrosis, excessive ROS production, increase in intracellular Ca 2+ , and osteogenic protein expressions could promote calcium deposition and osteogenesis transformation of A7R5 cells. As a result, vascular calcification occurred.
Calcium deposition occurs in two ways: the continuous growth of adhered HAP crystals on the cell surface 55,56 and the direct deposition of inorganic calcium and inorganic phosphorus components in the medium to the active sites on the damaged cell surface 57,58 . The culture medium contained 1.4 mmol/L inorganic phosphorus (KH 2 PO 4 ) and 2 mmol/L inorganic calcium (CaCl 2 ) 58 . However, their amounts were insufficient to form the amount of calcium phosphate deposition shown in Fig. 9. Therefore, the increase in calcium deposition was due to the continuous growth and deposition of inorganic calcium and phosphorus in the culture medium on the surface of the adhered HAP crystal.
Factors affecting the cytotoxicity of HAP crystals. The cytotoxicity of HAP crystals with different morphologies is not determined by a single factor. Cytotoxicity is affected by various physical parameters, including S BET , electrical conductivity, and Zeta of crystals. In addition, the cytotoxic effect of HAP is affected by cell membrane interaction, endocytosis, and intracellular Ca 2+ release.
1) The HAP toxicity was affected by S BET , conductivity, and Zeta of crystal  www.nature.com/scientificreports www.nature.com/scientificreports/ (c) c) Effect of the zeta potential of HAP The zeta potential values of HAP in pure water and culture medium were negative because of the abundance of the anionic P-OH group on the crystal surface 61 . Wilhelm et al. 62 suggested that the adsorption of negatively charged particles at positively charged sites via electrostatic interaction can lead to localized neutralization and subsequent bending of the membrane, thereby causing cellular uptake. In usual cases, the cell surface is weakly electronegative 63 . The toxicities of H-Sphere, H-Needle, and H-Rod crystals decreased as their negative zeta potential charge increased. The absolute zeta value of the H-Sphere crystal in the culture medium was the smallest, therefore, the highest affinity was observed in A7R5 cells, resulting in high cytotoxicity. H-Plate crystals had more negative charges, but their cytotoxicity was the greatest, which possibly because of their maximum S BET and high degree of endocytosis.
2) HAP toxicity was affected by cell membrane interaction, endocytosis, and intracellular Ca 2+ release. (d) a) The adhered HAP crystals would induce the breakdown of membrane lipids and the release of LDH (Fig. 3B), leading to cell membrane rupture. Severe cell membrane rupture would cause necrotic cell death (Fig. 8). www.nature.com/scientificreports www.nature.com/scientificreports/ (e) b) Nanocrystals were endocytosed by cells (Fig. 10). The endocytosed nanocrystals induced a decrease in Δψm (Fig. 7), destroyed lysosomal integrity (Fig. 6), and caused cell necrosis (Fig. 8). (f) c) The endocytosed HAP crystals caused a remarkable increase in intracellular Ca 2+ (Fig. 12). The sudden and intense release of Ca 2+ destroyed the osmotic pressure balance on lysosomal membranes 48 , causing excessive lysosomal disruption and cell necrosis (Fig. 8).
conclusions Four kinds of HAP nanoparticles with various shapes damaged the A7R5 cells to different degrees, resulting in decreased cell viability, disorganized cell morphology, disrupted cell membranes, increased intracellular ROS generation, decreased Δψm, decreased lysosome integrity, increased ALP, and increased intracellular calcium concentration, thereby leading to cell necrosis. The HAP-induced cytotoxicity showed the following trend: H-Plat e > H-Sphere > H-Needle > H-Rod. The nano-HAP with a large S BET , a high electrical conductivity, and a low Zeta elicited high cytotoxic effects. More calcium deposits on the cell surface, higher expression levels of osteogenic protein (BMP-2, Runx2, OCN, and ALP), and a stronger osteogenic transformation ability were observed in the crystal with a high cell cytotoxicity. This study could provide insights into the mechanism on how nano-HAPs injured vascular smooth muscle cells and induced vascular calcification. Figure 15. Schematic of the mechanism of cell injury and calcification caused by HAP nanoparticles with various shapes.