Supplementary Figure S1. Building up the geometric model of octapod iron oxide nanoparticles.

Transcription

1 Supplementary Figure S1. Building up the geometric model of octapod iron oxide nanoparticles. Geometric model of octapod iron oxide nanoparticles. (a, d, g) High-resolution transmission electron microscopy (HRTEM) images, (b, e, h) selected-area electron diffraction (SAED) patterns, and (c, f, i) geometric models of individual octapod iron oxide nanoparticles oriented along the [100], [111], and [110] directions, respectively.

2 Supplementary Figure S2. Characterization of octapod iron oxide nanoparticles. (a) X-ray powder diffraction (XRD) and (b) relevant SAED pattern of the as-prepared octapod iron oxide nanoparticles.

3 Supplementary Figure S3. Controlled synthesis of octapod iron oxide nanoparticles. TEM images of iron oxide nanoparticles obtained under the similar condition but different amount of NaCl: (a) 0 mg, (b) 2 mg, (c) 5 mg, and (d) 10 mg.

4 Supplementary Figure S4. Synthesis of iron oxide nanoparticles by other salts. TEM images of iron oxide nanoparticles synthesized by adding different salts: (a) no salts, (b) NaCl, (c) KBr, (d) NaF, (e) Na-oleate, and (f) NaOH.

5 Supplementary Figure S5. Synthesis of iron oxide nanoparticles by adding other chloride sources. TEM images of iron oxide nanoparticles prepared by replacing NaCl with (a) CTAC, (b) CTAB, (c) KCl, and (d) KBr.

6 Supplementary Figure S6. Analysis of octapod iron oxide nanoparticles. The X-ray spectroscopy (EDS) and X-ray photoelectron spectroscopy (XPS) analysis of octapod iron oxide nanoparticles showed the evidence of chloride ions on the octapod iron oxide nanoparticles. (a) Geometric models illustrate the Cl - ions adsorbed on the surface of octapod iron oxide. (b) Crystal structure model corresponding to the [311] facets of the octapod iron oxide nanoparticles. EDS analysis of (c) the region of accumulative octapod iron oxide nanoparticles and (d) blank region in TEM sample. XPS analysis of octapod iron oxide nanoparticles showed that (e) the Fe 2p doublet at ev (Fe 2p3/2) and ev (Fe 2p1/2) corresponds to Fe 3 O 4 and (f) the Cl 2p3/2 peak 43 at ev corresponds to the binging energy of Cl-Fe 3+. The red line is a fitted result of spectrum.

7 Supplementary Figure S7. Models for calculation. (a) The real geometric model. (b) The simplified geometric model. (c) Schematic of the R and r. R corresponding to the simulated spherical ball which covering the full octapod nanoparticle and the r corresponding the spherical nanoparticle with equal geometric volume to octapod nanoparticle.

8 Supplementary Figure S8. Magnetic property of octapod iron oxide nanoparticle. (a) Zero-Field-Cooled/Field-Cooled (ZFC/FC) curves and (b) M-H curves (low magnetic field areas at 300 K) of Octapod-14, Octapod-20, Octapod-30, and Octapod-36, respectively.

9 Supplementary Figure S9. TEM and DLS data of water-dispersible iron oxide nanoparticles. TEM images and dynamic light scattering (DLS) analysis of water-dispersible (a, b, c) Octapod-30, (d, e, f) Spherical-16, (g, h, i) Octapod-20, and (j, k, l) Spherical-10 nanoparticles, respectively, (insets: optical photos).

10 Supplementary Figure S10. MR relaxivities of the iron oxide nanoparticles. The analysis of relaxation rate R 2 (1/T 2 ) vs Fe concentration for Octapod-30, Octapod-20, Spherical-16, and Spherical-10 samples.

11 Supplementary Figure S11. The comparison of Spherical-25 and Octapod-20. TEM images of (a) Spherical-25 and (b) Octapod-20. (c) DLS data suggest that the Spherical-25 and Octapod-20 have the similar hydrodynamic diameters. (d) M-H curves of Octapod-20 and Spherical-25 measured at 300 K (inset: M-H curves of Octapod-20 and Spherical-25 in low magnetic field areas). (e, f) Comparison of r 2 values of Octapod-20 and Spherical-25. The error bars represent standard deviation (± s.d.) of five independent experiments.

12 Supplementary Figure S12. In vitro cytotoxicity of octapod iron oxide nanoparticles. MTT assay of HepG2 cells incubated with various concentration of octapod iron oxide nanoparticles for 24 h (n = 3/group, ± s.d.).

13 Supplementary Figure S13. Prussian blue staining. Prussian blue staining images of liver tissues from (a) BALB/c mice and (b) nude mice at 1 h after intravenous injection of Octapod-30 and Spherical-16. The results indicated that the accumulation amount of Octapod-30 is comparable to that of Spherical-16 after administration of the same injection dose. Moreover, the accumulation amount decreased obviously when we reduced the injection dose of Octapod-30 to 0.5 mg/kg. Scale bar: 50 m for all images.

14 Supplementary Figure S14. Liver uptake of iron oxide nanoparticles. The ICP-MS analysis of liver uptake of Spherical-16 and Octapod-30 (n = 3/group, ± s.d.).

15 Supplementary Figure S15. In vivo MR imaging. (a) In vivo MR images of mice at 0, 0.5, and 1 h after intravenous injection of Octapod-30 (0.5 mg Fe/kg, left) and Spherical-16 (1 mg Fe/kg, right) in transverse plane, respectively. (b) Quantification of liver contrast collected at 0, 0.5, and 1 h after administration of Octapod-30 and Spherical-16 in BALB/c mice (n = 3/group, ± s.d.).

16 Supplementary Figure S16. Induced magnetic fields around the iron oxide nanoparticles. (a) Illustration of the models for calculating the intensity of induced magnetic fields away from the spherical and octapod iron oxide nanoparticles. The left and right sides are divided by the pole for spherical nanoparticle and the median line of the plane (green line) for octapod nanoparticles. Simulation of the induced magnetic field lines around (b) octapod iron oxide nanoparticles and (c) spherical iron oxide nanoparticles. The x = 3.75, 6.25, and 9.75 means the magnetic field lines are 3.75, 6.25, and 9.75 nm away from the surface of nanoparticle, respectively.

17 Supplementary Table S1. MR signal-to-noise ratio (SNR) changes of region of interest (ROI) before and after intravenous injection of Octapod-30 and Spherical-16, respectively (n = 3/group). We calculated the signal-to-noise ratio (SNR) by the equation: SNR liver = SI liver / SD noise, where SI represent signal intensity and SD represent standard deviation. Then we calculated the SNR changes of ROI by the equation: SNR = SNRpost - SNRpre / SNRpre. Octapod-30 (1mg /kg) Spherical-16 (1mg /kg) Octapod-30 (0.5mg /kg) SNR pre (%) SNR 0.5 h (%) 36.1 ± ± ± 2.5 SNR 0.5 h (%) 63.9 ± ± ± 2.5 SNR 1 h (%) 32.7 ± ± ± 1.9 SNR 1 h (%) 67.3 ± ± ± 1.9

18 Supplementary Table S2. MR contrast-to-noise ratio (CNR) changes of tumor-to-liver contrast pre- and post- injection of Octapod-30 and Spherical-16 with a dose of 2.0 mg/kg, respectively (n = 3/group). To qualify the efficacy of the enhancement, we introduced the contrast-to-noise ratio (CNR), which was given by CNR = (SNR tumor SNR liver ) / SNR tumor, Where the SNR was defined previously. Then we calculated the CNR changes of ROI by the equation: CNR = CNRpost/CNRpre 1. CNRpost/CNRpre (0 h) (%) CNRpost/CNRpre (0.5 h) (%) CNR 0.5 h (%) CNRpost/CNRpre (1 h) (%) CNR 1 h (%) CNRpost/CNRpre (2 h) (%) CNR 2 h (%) CNRpost/CNRpre (4 h) (%) CNR 4 h (%) Spherical ± ± ± ± ± ± ± ± 2.7 Octapod ± ± ± ± ± ± ± ± 8.5

19 Supplementary Table S3. Comparison of the induced magnetic field intensity away from the surface of nanoparticles with the same distance (n = 3/group). We calculated the intensity of stray field generated by octapod and spherical iron oxide nanoparticles, and compared the field intensity at various distances from the surface of particles. Intensity of the magnetic field (Oe) Distance to the Octapod-30 Spherical-16 surface (nm) left right left right ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± (3.93 ± 0.2) 10-3 (3.93 ± 0.2) 10-3 (1.43 ± 0.1) 10-3 (1.43 ± 0.1) 10-3

20 Supplementary Table S4. Properties of T 2 contrast agents. Summary of various T 2 contrast agents with different materials, core sizes, morphologies, hydrodynamic diameters, r 2 relaxivities, and magnetization values. Materials Core size (nm) Shape Hydrodynamic diameter (nm) r 2 (mm -1 S -1 ) Field (T) M s (emu/g) (Zn x Fe 1-x )Fe 2 O 4 nanoparticles 15 Spherical MnFe 2 O 4 nanoparticles 12 Spherical N/A (Zn x Mn 1-x )Fe 2 O 4 nanoparticles 15 Spherical FeCo/graphiticnanocrystals 7 Spherical N/A Fe 3 O micell 16 Spherical 110 ± N/A 35 Iron/iron oxide nanoparticles 16 ± 1.5 Spherical N/A Iron@MnFe 2 O 4 nanoparticles 16 Spherical Octapod iron This oxide 30 ± 3 Octapod 58 ± work nanoparticles N/A: Not Available. Refs

21 Supplementary Note 1. Calculation of the conversion factor κ. Nomenclature γ proton gyromagnetic ratio L thickness of an impermeable surface coating V * N volume fraction Fe3O4 the number of Fe 3 O 4 nanoparticles in the sample M s saturation magnetization V sample volume of the sample r effective radius of magnetic V Fe3O4 volume of the Fe 3 O 4 nanostructure nanoparticles D diffusivity of water molecules Q Fe quality of the Fe ions in iron oxide nanoparticle ρ Fe3O4 density of the Fe 3 O 4 nanoparticles M Fe molar mass of Fe m Fe mass of the Fe ions in the sample κ conversion factor In the motional average regime (MAR), the relaxation rate R 2 is given by (assuming that the nanoparticles are spherical): 1/T 2 = R 2 = (256π 2 γ 2 /405) V * 2 M s r 2 / D (1+L/r) (S1) Where V * is the volume fraction, which is defined as 46 : V * = N Fe3O4 (4πr 3 /3)/V sample (S2) For the spherical nanoparticles, V Fe3O4 = 4πr 3 /3, so we can transform (S2) to: V * = N Fe3O4 V Fe3O4 /V sample. (S3) For the Fe 3 O 4 nanoparticles, the mass of the Fe ions can be calculated by: Q Fe = 168/232 ρ Fe3O4 V Fe3O4 (S4) Thus, we can calculate the mass of Fe in any sample by the equation (S4): m Fe = Q Fe N Fe3O4 (S5) The C Fe is defined as: C Fe = m Fe / M Fe V sample (S6) Based on these, we can transform V * to C Fe by a conversion factor κ, which can be calculated by κ = 1.38 M Fe /ρ Fe3O4.

22 Supplementary Note 2. Detailed calculations on the structure of octapod iron oxide nanoparticles. The calculations about the volume of the octapod nanoparticles are showed as follows (we simplify the octapod structure as the Supplementary Fig. S7b): As shown in the simplified geometric model (Supplementary Fig. S7b), the octapod model is composed of 8 tetrahedrons, 4 pyramids, and 1 cube. For simplicity, we donated that L ef = α, L ae = β, L gh = γ, L aj = δ, and L io = ε. By calculating from the model, we got that β = 3 α, γ = 2 α, δ = 4α, and ε = 2 α. For the tetrahedron, L ae = L ag = L ac = L ce = α and L ae = β. h tet = L ab = β (2/3( α 1/4α )), V tet = 1/3 3/2α α 1/2 h tet = 3 2 /6α For the pyramid, L gh = c= 2 α, L ge = L gc = L ec = L hc = α. h pyr = α (1/2( γ α )), V pyr = 1/3 α γ h pyr = 3 2 /6α For the cube, L gh = 2 α, L eg = L ef = α. V cub = 2 α α α = 3 2 α So V octapod = 8V tet + 4V pyr + V cub = 3 2 /6α /6α α, V spherical = 4/3π r 3, When the geometric volumes of octapod particle and spherical particle are the same, 3 2 /6α /6α α = 4/3π r 3, r 1.01α. We then compared the areas of octapod and sphere under the same geometric volumes. S octapod = 24S aeg + 6S egfm = 11 /2α α 2 = α 2 + 8α 2. For the sphere, S spherical = 4πr 2 (r =1.01α). So S octapod 3.73 S spherical, means that the surface-to-volume (S/V) ratio of octapod nanoparticle is 3.73 times as high as that of spherical nanoparticle. For the octapod nanoparticle, R = 2 2 ( δ/2) ε = 6 α 2.45α, which means R 2.42r under the same geometric volumes (Supplementary Fig. S7c ). In our case, the efficient diameters (2R) of Octapod-30 and Octapod-20 are 40 ± 2 and 26 ± 1 nm, respectively. So we chose the spherical iron oxide nanoparticles with diameters (2r) of about 16 nm and 10 nm accordingly as control samples for comparison.

23 Supplementary Methods Reagent. FeCl 3 (99%), NaCl (AR), KCl (AR), KBr (AR), NaF (AR), hexane, sodium oleate, isopropanol, and ethanol were purchased from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China). 1-octadecene (90%), oleic acid (90%), hexadecyl trimethyl ammonium Chloride (CTAC), and hexadecyl trimethyl ammonium Bromide (CTAB ) were purchased from Alfa Aesar. All reagents were used as received without further purification. Characterization. Transmission electron microscopy images (TEM) were taken on JEOL JEM-2100 at 200 kv. The structures of the nanoparticles were characterized by X-ray powder diffraction (Panalytical X-pert diffractometer with Cu-Ka radiation). The hydrodynamic diameters of nanoparticles were measured with a particle size analyzer (DLS). M-H curves were obtained by the superconducting quantum interference device (SQUID). The iron concentrations of nanoparticles were measured with inductively coupled plasma atomic emission spectroscopy (ICP-AES). Synthesis of iron oleate complex. The synthesis of iron oleate was carried out according the published procedure with minor modification 16. In a typical experiment, 4.56 g of sodium oleate (15 mmol) and 0.81g of FeCl 3 (5 mmol) were dissolved in a mixture of 20 ml of distilled water and 10 ml of ethanol. The resulting solution was heated to 70 C and kept at that temperature for 4 h under argon atmosphere. When the reaction was completed, the iron oleate was washed three times with distilled water in a separatory funnel. After hexane was volatilized naturally in the dish, we obtained the iron oleate complex in a waxy solid form, and stored it at room temperature. Synthesis of spherical iron oxide nanoparticles with a size of 16 nm. In a typical experiment, 0.93 g iron-oleate (1 mmol) synthesized as described above and 160 l oleic acid (0.5 mmol) were dissolved in 15 ml 1-octadecene at room temperature. The mixture was degassed in vacuum for 30 min and backfilled with argon to remove any low volatile impurities and oxygen at room temperature. And then, we heated the reaction solution to 320 C with a constant heating rate of 3.3 C min -1, and kept at that temperature for 1 h. The initial reddish-brown color of reaction solution turned brownish-black. The

24 resultant solution was then cooled to room temperature and mixed with 30 ml isopropanol to precipitate the nanoparticles. The nanoparticles were separated by centrifugation and washed 3 times with ethanol. After washing, the nanoparticles were dissolved in hexane for long term storage at 4 C. Synthesis of spherical iron oxide nanoparticles with a size of 25 nm. In a typical experiment, 0.93 g iron-oleate (1 mmol) and 160 l oleic acid (0.5 mmol) were dissolved in 10 ml trioctylamine at room temperature. The mixture was degassed at room temperature in vacuum for 30 min and backfilled with argon to remove any low volatile impurities and oxygen. And then, we heated the reaction solution to 350 C with a constant heating rate of 3.3 C min -1, and kept at that temperature for 5 h. The initial reddish-brown color of the reaction solution turned into brownish-black. The resulting solution was then cooled to room temperature and mixed with 30 ml ethanol to precipitate the nanoparticles. The nanoparticles were separated by centrifugation and washed 3 times with ethanol. After washing, the nanoparticles were dissolved in hexane for long term storage at 4 C. Synthesis of octapod iron oxide nanoparticles with the edge length of 30 nm. In a typical synthesis of octapod iron oxide nanoparticles, iron oleate (0.8 g, 0.86 mmol), NaCl (10 mg, 0.17 mmol), oleic acid (110 L, 0.35 mmol), and distilled water (60 L) were mixed together with 10 ml of 1-octadecene. The resulting solution was degassed in vacuum for 30 min and backfilled with argon to remove any low volatile impurities and oxygen at room temperature. And then, we heated the reaction solution to 320 C with a constant heating rate of 3.3 C min -1, and kept at the temperature for 2 h. The color of the solution changed from reddish-brown to transparent orange and finally brownish-black. Then the solution was cooled to room temperature and mixed with 30 ml of isopropanol to precipitate the nanoparticles. The nanoparticles were separated by centrifugation and washed 3 times with ethanol. The final product was dissolved in hexane for long-term storage at 4 C. Synthesis of octapod iron oxide nanoparticles with the edge length of 20 nm. The procedure of synthesis of Octapod-20 nanoparticles is similar to that of Octapod-30. After heating to 320 C, the

25 solution was kept at that temperature for 1 h. The nanoparticles were separated by centrifugation and washed 3 times with ethanol. After washing, the nanoparticles were dissolved in hexane for long term storage at 4 C. Preparation of water soluble HDA-G 2 encapsulated nanoparticles. Typically, 1 ml of chloroform containing 20 mg of HDA-G 2 was added to 1 ml of chloroform containing 10 mg of nanoparticles, the container was left to open in a fume hood to evaporate the solvent slowly at room temperature. The residual chloroform was removed completely by pump, then the dry sample was re-dispersed in water by sonication. Further purification of the water-dispersible sample was performed by size exclusion chromatography (PD-10 column, GE Healthcare Life Science). The final aqueous solution was stored at 4 C for further use. Measurement of MR relaxivity of iron oxide nanoparticles. To measure the T 2 relaxivity, Octapod-30, Octapod-20, Spherical-16, and Spherical-10 with different iron concentrations were dispersed in 1% agarose solution. The samples were scanned using a multi-echo T 2 -weighted fast spin echo imaging sequence (TR/TE=2000/20, 40, 60, 80, 100 ms, slice thickness = 2 mm) by a 7 T MRI scanner (Varian 7 T micro MRI System). Cell culture. The HepG2 cells were purchased from Cell Bank of Chinese Academy of Sciences (Shanghai, China) and cultured in Dulbecco s Modified Eagle s Medium (DMEM medium) supplemented with 10% fetal bovine serum (FBS, Hyclone) and antibiotics (100mg/mL streptomycin and 100 U/mL penicillin). All cells maintained in a humidified atmosphere of 5% CO 2 at 37 C. In vitro cytotoxicity evaluation. The HepG2 cells (1 x 10 4 ) were seeded in 96-well plates and incubated for 12 h in DMEM (containing 10% FBS). After washed cells twice with PBS, we added fresh medium containing octapod iron oxide nanoparticles with different concentrations (the equivalent Fe concentrations were 100, 67, 44, 30, 20, 13, 9, 6, 4, and 0 g Fe/mL) and incubated the cells for 24 h. Each experiment in the same concentration was performed in five wells. We replaced the growth

26 medium with DMEM containing 0.5 mg/ml of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and incubated for another 4 h. After discarding the culture medium, 100 L of DMSO was added to dissolve the precipitates and the resulting solution was measured for absorbance at 492 nm using a MultiSkan FC microplate reader (Thermo scientific). Prussian blue staining. After MR imaging, we sacrificed the mice and kept the liver of mice in the optimal-cutting-temperature (O.C.T) compound and stored at -80 C. When the mixture was freezing, we cut the samples into 10 m thick slices and fixed with ice-cold acetone for 5 min immediately. After drying at room temperature for 5 min, we put the slide into the staining solution (20% hydrochloric acid and 10% potassium ferrocyanide solution mixture, 1:1 volume ratio) for 30 min, and counterstained with eosin for 5 min. Then, we used 90%, 95%, and pure water to clear the slide for 3 times. Calculating the induced magnetic fields around the iron oxide nanoparticles. We carried out 3D numerical modeling by solving the Landau-Lifshitz-Gilbert (LLG) equation, which is viable for studying the magnetization process on nanoscopic magnets. We chose Octapod-30 and Spherical-16 as examples with the following parameters: M s = 71 emu/g for Octapod-30 and 65 emu/g for Spherical-16; external static field is set at 7 T; exchange stiffness constant A=1.0e -6 erg/cm; Gilbert damping constant is 1.0; and the unit cell dimensions are 1 nm 1 nm 1 nm. The magnetic field data from LLG could be dealt with Matlab to show the magnetic field distribution outside the particle. We calculated the intensity of stray field generated by octapod and spherical iron oxide nanoparticles, and compared the field intensity at various distances from the surface of particles.

27 Supplementary References 43. Wagner, C.D., Moulder, J.F., Davis, L.E. & Riggs, W.M. Handbook of X-ray photoelectron spectroscopy. (Perking-Elmer Corporation, 1979). 44. Seo, W.S. et al. FeCo/graphitic-shell nanocrystals as advanced magnetic-resonance-imaging and near-infrared agents. Nat. Mater. 5, (2006). 45. Cheong, S. et al. Simple synthesis and functionalization of iron nanoparticles for magnetic resonance imaging. Angew. Chem. Int. Ed. 50, (2011). 46. Vuong, Q.L., Gillis, P. & Gossuin, Y. Monte Carlo simulation and theory of proton NMR transverse relaxation induced by aggregation of magnetic particles used as MRI contrast agents. J. Magn. Reson. 212, (2011).