Core-shell NaGdF4@CaCO3 nanoparticles for enhanced magnetic resonance/ultrasonic dual-modal imaging via tumor acidic micro-enviroment triggering

For cancer diagnosis, a paramount challenge still exists in the exploring of methods that can precisely discriminate tumor tissues from their surrounding healthy tissues with a high target-to-background signal ratio. Here, we report a NaGdF4@CaCO3-PEG core-shell nanoparticle which has the tumor acidic microenvironment enhanced imaging signals of ultrasound and magnetic resonance. Under the acidic conditions, the CaCO3 shell will gradually dissolve which then facilitate the interaction of NaGdF4 with the external aqueous environment to enhance water proton relaxation. Meanwhile, the CO2 bubbles generated by the CaCO3 dissolvement will generate strong elastic echo for US detection. The core-shell structure of NaGdF4@CaCO3-PEG can be observed by TEM, and its composition can be determined by STEM. The acid triggered generation of CO2 bubbles and the enhancement of MRI signal could be demonstrated in vitro, and the excellent dual-modal magnetic resonance/ultrasonic cancer imaging abilities of NaGdF4@CaCO3-PEG could be also proved at the tumor site in vivo. The here described proof-of-concept nanoparticles with pH triggered magnetic resonance/ultrasonic dual-modal imaging enhancement, may serve as a useful guide to develop various molecular imaging strategies for cancer diagnosis in the future.

the widely used paramagnetic gadolinium ions (Gd-(DTPA)] −2 (Magnevist) and [Gd-(DOTA)] −1 (Dotarem) [21][22][23] , or gas-filled echogenic microbubbles 24,25 , that are always "on", emitting constant signals regardless of their proximity or interaction with target tissues, cells, or environmental markers. As a result, a large volume of nonspecific signal which might lead to a poor signal-to-noise ratio, makes the anatomical features of interested tissue difficult to distinguish. A more attractive contrast agent whose signal should be switched from OFF to ON in response to specific biological stimulus, which will further maximize the signals of targets and minimize the background signals, which in turn could improve the sensitivity and specificity [26][27][28][29] .
Several activatable CAs that respond to tumour-related factors, such as pH and redox potential, have been developed for tumour-specific MRI or US. For example, Mi et al. have reported a MRI contrast agent that rapidly amplify the magnetic resonance signals in response to pH via releasing confined Mn 2+ ions from pH-sensitive calcium phosphate (CaP) nanoparticles to the aqueous environment 30 . Min et al. developed the calcium carbonate (CaCO 3 ) nanopartciles which exhibited strong echogenic signals at tumoral acid pH by producing CO 2 bubbles and showed excellent echo persistence 10 . Even so, some intrinsic drawbacks of US and MRI still cannot be avoided, for example, US has a poor tissue discrimination ability while MRI cannot provide real-time images and usually time consuming. Thus, developing a sensitive dual modal imaging (US and MRI) agent would not only make them as favorable tools for precisely visualizing biological and physiological changes with high signal-to-noise ratio, but also would render synergistic efficacy to overcome their own inherent limitations 31 .
As a MRI contrast agent, various size of NaGdF 4 nanoparticles (NPs) with well-defined size distributions could be readily synthesized via pyrolysis methods [32][33][34][35] . Their MRI performances have been demonstrated to increase with the decreasing of the NPs' size, attributed to the increased number of surface Gd 3+ ions relative to the core ions. Therefore, it will be very useful to synthesize ultrasmall NaGdF 4 NPs, e.g., smaller than 5 nm, to provide better MRI signals 6 . Furthermore, compared with the Gd-(DTPA) or Gd-(DOTA), the renal clearable Gd-based NPs would have better biosafty since it could not induce the excessive Gd 3+ ion leakage to cause biological toxicity 36,37 , therefore it might be a more promising candidate for disease diagnosis. As for a new type of US contrast agent, CaCO 3 nanoparticles with rigid structure can penetrate into host tumoral environments for cancer imaging, while the frequently used gas-filled microbubbles suffer from inherent drawbacks, such as low stability, short half-life in blood and low penetration ability due to the large size, which is limited to the imaging of intravascular structures.
Herein, we designed a pH-responsive nanoparticle which could significantly enhance the contrast of MRI and US signals in tumor as illustrated in Fig. 1. The shell of CaCO 3 was deposited onto the core surface of NaGdF 4 nanoparticles through the microemulsion method; in addition, physicochemical properties of nanoscale systems, such as size, dispersibility, and toxicity were systematically analyzed; furthermore, the US and MRI imaging enhanced efficiency were both evaluated in vitro and in vivo, which clearly proved that our probe could be utilized for sensitive and specific tumor imaging with responding to extracellular acidic microenvironments.

Results and Discussion
Characterization of NaGdF 4 @CaCO 3 -PEG. As shown in Fig. 1, monodispersed spherical NaGdF 4 nanocrystals were chosen as the morphology-deciding template to obtain magnetic nanoparticles, which were fabricated by thermal decomposition method using rare earth perchlorate as the RE precursor. From TEM images ( Fig. 2A), the synthesized NaGdF 4 exhibit a very well particle size distribution around 8~10 nm, with a regular spherical morphology. A thin layer of dense CaCO 3 was deposited onto the surface of NaGdF 4 to form NaGdF 4 @ CaCO 3 core/shell nanoparticles by the well-known microemulsion method, which could further transfer the hydrophobic NaGdF 4 nanoparticle into aqueous soltion. In detail, the CaCl 2 aqueous solution was first added into the cyclohexane solution of hydrophobic NaGdF 4 , while triton X-100, 1-hexanol were used as nonionic surfactants to create a stable water-in-oil emulsion after vigorous stirring. At this stage, the hydrophobic NaGdF 4 will transfer from oil phase to water droplets as the surfactants can self-assemble on these nanoparticles. Meanwhile, the hydrophilic groups (hydroxyl) of the surfactants has a well affinity with Ca 2+ which could recruit these ions on the surfaces of the NaGdF 4 nanoparticles. After adding Na 2 CO 3 aqueous solution, the precipitation reaction between Ca 2+ and CO 3 2− therefore produces a layer of CaCO 3 on the NaGdF 4 nanoparticles. The obtained nanoparticles show a size of 10~12 nm, as well as with aspherical-like morphology. Then, PEG 8000 was adsorbed on the surface of NaGdF 4 @CaCO 3 nanoparticles through Van der Waals' force, which could award good water dispersibility. Element mapping of NaGdF 4 @CaCO 3 -PEG with a thin or thick shell all shows the co-existence of Gd, F and Ca elements ( Figure S1), which exhibit a much higher Ca signal compared to the core only NaGdF 4 nanoparticles. However, the Ca signal on the surface of NaGdF 4 is very weak on the outlayer of NaGdF 4 core due to the very small size of NaGdF 4 core (10 nm) and the very thin CaCO 3 coating layer (2 nm). Nevertheless, we still ensure that the CaCO 3 is on the shell because this layer is amorphous, while the NaGdF 4 core is highly crystal with noticeable lattice, as shown in the enlarged image inserted in Fig. 2B. In addition, the average size of the NaGdF 4 in cyclohexane was further determined to be 8 ± 2 nm by the DLS experiments ( Figure S2), which is consistent with the TEM results. After depositon of CaCO 3 , the average hydrodynamic size of the NaGdF 4 @CaCO 3 and NaGdF 4 @CaCO 3 -PEG in water were increased to 100 ± 10 and 120 ± 15 nm, respectively, which may be due to the formation of a hydration layer on the surface of nanoparticles or the formation of small aggregations. Furthermore, the successful surface PEGlyation of NaGdF 4 @CaCO 3 was also confirmed by FT-IR spectroscopy (Fig. 3). For analyzing the spectrum of PEG 8000 , the peaks at around 2917 cm −1 and 1108 cm −1 were observed corresponding to the CH 2 stretching and vibration groups. These two peaks could be also observed in the spectra of NaGdF 4 @CaCO 3 -PEG, in addition to the typical peaks of NaGdF 4 @CaCO 3 at 502 cm −1 . These results suggest that the PEG 8000 was successfully modified on the surface of NaGdF 4 @CaCO 3 . To make the CaCO 3 coating more evident form the IR spectra, we re-fabricated another NaGdF 4 @CaCO 3 with a thicker shell via increasing the feeding amount of CaCl 2 , and analysis their surface groups through FT-IR measurement. The characteristic IR absorption peaks of CaCO 3 at 1430, 876, 712 cm −1 can be clearly presented in thick layered NaGdF 4 @CaCO 3 ( Figure S3).
Before further bio-imaging applications, it is necessary to determine the cytotoxicity of these nanoparticles. To assess the in vitro cytotoxicity of NaGdF 4 @CaCO 3 -PEG, the standard Cell Counting Kit (CCK-8) assay and live/dead staining were conducted by using LN3 and NIH3T3 cell lines. As shown in Figure S4, cell viability is not affected by the NaGdF 4 @CaCO 3 -PEG in the concentration ranged from 0 to 400 μg•mL −1 , which suggested that the NaGdF 4 @CaCO 3 -PEG has low cytotoxicity so that it could be further applied for bio-imaging.
In Vitro Ultrasound Contrast Enhancement Ability of NaGdF 4 @CaCO 3 -PEG. First, the characteristics of CO 2 gas generation by NaGdF 4 @CaCO 3 -PEG was studied through detecting the bubble generation from the particles in PBS buffer with different pH values (pH 7.4, 7.0, 6.8 and 5.0). As shown in Fig. 4, initially, NaGdF 4 @CaCO 3 -PEG nanoparticles generate few bubbles at pH 7.4, 7.0 and 6.8, but significantly more bubble generation could be observed at pH 5.0. The degree of bubble generation continuously decreased at pH 5.0 over time, but the bubble generation was first increased and then decreased at pH 7.4, 7.0 and 6.8. The number of bubbles in pH 5.0 was counted to ca. 119, which was much higher than that in pH 7.4 (almost no bubbles), after 1 min incubation. However, with increasing the time, the number of bubbles was dramatically decreased at pH 5.0, and only 2 bubbles could be observed after 60 mins incubation ( Figure S5). It is noteworthy that bubble generation at lower pH was higher than that at slightly stronger pH at all-time points. This result suggested that the NaGdF 4 @ CaCO 3 -PEG might enhance US signal in the low pH environment by gas generation, since CaCO 3 could generate more CO 2 gas in an acidic environment by improved dissolution, which would be beneficial for imaging the tumor due to a little acidic microenvironment of the tumor micro-environment.
Next, we measured the in vitro echogenic performance of the NaGdF 4 @CaCO 3 -PEG as a function of pH using an aqueous dispersion test at 37 °C. As shown in Fig. 5, the NaGdF 4 @CaCO 3 -PEG at pH 7.4 showed no significant contrast signals under a US field, most likely because the NaGdF 4 @CaCO 3 -PEG did not release sufficient CO 2 bubbles for echogenic reflection. This result is well consistent with the results of the pH-dependent CO 2 generation (Fig. 4). In contrast, US contrast images from the NaGdF 4 @CaCO 3 -PEG showed first increased and then decreased signal at all-time points at weakly acidic pH conditions (pH 7.0 and 6.8). This is a well understood phenomenon in which the enhanced echo intensity at weakly pH levels is ascribed to the relaxedly facilitated formation of CO 2 bubbles. Due to faster dissolution of CaCO 3 phases at pH 5.0, the echo improved immediately and faded away earlier than at higher pH. The excellent echo persistence at weak acidic pH can be ascribed to the relaxedly ionization of CaCO 3 solid phases. The contrast enhancement that derives from the NaGdF 4 @ CaCO 3 -PEG at lower pH suggested that generation of CO 2 bubbles was responsible for echogenic US resonation. US imaging results were further quantitatively analyzed using region-of-interest (ROI) gray value quantification ( Figure S6A). The gray value in pH 5.0 and pH 7.4 was respectively determined to be 139.2 ± 10.7 and 14.2 ± 1.2, after 30 min incubation. This result is in accordance with the CO 2 bubbles generation. Meanwhile, in order to further confirm that the US signal was derived from the dissolution of CaCO 3 at acidic solution rather than the NaGdF 4 itself, we tested US signal of the core-only (NaGdF 4 ) nanoparticle at pH 5.0 and 7.4, and the results are shown in Figure S7. NaGdF 4 nanoparticle showed no US signal at both pH 5.0 and 7.4, due to no CO 2 gas generation from NaGdF 4 nanoparticle. The feature of gas-generating at a broad range of weak acidic pH would be propitious to US imaging of tumors, which have tissue heterogeneity and diverse acidic pH levels.   Scientific RepoRts | 7: 5370 | DOI:10.1038/s41598-017-05395-w pH-Dependent In Vitro MRI Properties of NaGdF 4 @CaCO3-PEG. Since the doping of Gd 3+ -ions, NaGdF 4 @CaCO 3 -PEG could act as a T 1 MRI contrast agent as well. T 1 -weighted MR images of the NaGdF 4 @ CaCO 3 -PEG showed enhancing signal intensity when Gd 3+ concentrations were increased. Firstly, the T 1 -weighted MR image of the NaGdF 4 @CaCO 3 -PEG was studied at different pH conditions. As shown in Fig. 6, the NaGdF 4 @CaCO 3 -PEG at pH 7.4 showed no significant enhanced contrast signals under a MRI field, most likely because the shell of CaCO 3 did not decompose so that water molecular could not access to NaGdF 4 core. In contrast, MRI contrast images from the NaGdF 4 @CaCO 3 -PEG were slightly enhanced at weakly acidic pH conditions (pH 7.0, 6.8). It is noteworthy that we demonstrated strongest MRI contrast images at pH 5.0, which might due to the decomposition of the CaCO 3 shell so that water molecular could access to the NaGdF 4 core. To further demonstrate this mechanism, we compared the T 1 -MRI of NaGdF 4 @CaCO 3 -PEG to that of core-only (NaGdF 4 ) nanoparticles at pH 7.4, and the results are shown in Figure S6. Of course, NaGdF 4 showed a much brighter MR image than NaGdF 4 @CaCO 3 -PEG at pH 7.4. With the increasing of incubation time, the contrast of MRI images gradually enhanced at pH 7.2 and 6.8, but almost no change at pH 7.4. MRI imaging results were then quantitatively analyzed using region-of-interest (ROI) gray value quantification ( Figure S6B). The gray value in pH 5.0 and pH 7.4 was respectively determined to be 182.2 ± 10.1 and 79.9 ± 3.1, after 20 min incubation. These data are very consistent with the theory that the contrast of T 1 MRI imaging is closely related with water accessibility, and the enhancement of contrast associated with sufficient water contacting.
In addition, the T 1 value of the NaGdF 4 @CaCO 3 -PEG was evaluated by using a 0.5 T MRI scanner to determine whether the pH can influence the T 1 -weighted MR imaging performance (Fig. 7). As shown in Fig. 7A, the longitudinal (T 1 ) were measured at 0.5 T magnetic field based on the Gd concentration of NaGdF 4 @CaCO 3 -PEG at pH 7.4 and 5.0, and the relaxivities were determined to be 0.42 mM −1 •s −1 and 1.64 mM −1 ·s −1 , respectively. At pH 5.0, it showed about 4 folds higher relaxivity than that at pH 7.4, which further proofed that the T 1 -weighted MR imaging enhancement of the NaGdF 4 @CaCO 3 -PEG was influenced by pH levels. At the same time, Fig. 7B shows the magnetic resonance signal enhancing capability of the NaGdF 4 @CaCO 3 -PEG nanoparticles as a function of Gd concentration ranging from 1.25 to 12.5 mM. Compared with water (Gd, 0 mM), the measured T 1 -weighted image contrast gradually increased with the increasing of Gd concentrations.

In Vivo US Imaging of Tumor with NaGdF 4 @CaCO3-PEG.
To verify the potential of NaGdF 4 @ CaCO 3 -PEG for US imaging of tumor, we executed an intra-tumor injection of NaGdF 4 @CaCO 3 -PEG dispersion into LN3 tumor-xenograft-bearing nude mice and monitored the US images as a function of time (Fig. 8). After injection right away, we couldn't acquire any contrast enhancment of the US signal. It is gratifying that enhancement of US signals was obtained after 1 h of injection, and this contrast enhancement maintained more than 2 h. In addition, this phenomenon is very consistent with the in vitro experiments. US imaging results were then quantitatively analyzed using region-of-interest (ROI) gray value quantification. As shown in Fig. 8B, the gray value of tumor site was increased from 32.5 ± 3.9 (0 h post-injection) to 53.1 ± 7.8 (1 h post-injection), and still maintained at 38.7 ± 1.35 after 2 h injection. Therefore, we can conclude that the NaGdF 4 @CaCO 3 -PEG could generate bubbles in tumor tissues then produce sufficient echogenic reflectivity under a US field.
In Vivo MR Imaging of Tumor with using NaGdF 4 @CaCO 3 -PEG. NaGdF 4 @CaCO 3 -PEG nanoparticles which have the advantages of highly efficient of T 1 contrast enhancement ability may hold great promise to serve as a novel MRI contrast enhancing agent. Therefore, we performed MRI study on tumor-xenograft-bearing nude mice at a 7.0 T clinical scanner by intratumor injection of NaGdF 4 @CaCO 3 -PEG with a dose of 200 μL of 2 mg•mL −1 . Right after the injection, the MRI contrast enhancement was not found at tumor site; however, after one hour of intratumor injection, we were delighted to find contrast enhancement of MRI signal and the signal strength remained unchanged over time up to 3 h (Fig. 9A). T 1 imaging results were then quantitatively analyzed using region-of-interest (ROI) quantification. As shown in Fig. 9B, the gray value of tumor site was increased from 102 (0 h post-injection) to 153 (3 h post-injection). Overall, these results suggested that the NaGdF 4 @ CaCO 3 -PEG could act as a potential MRI contrast enhancing agent for tumor imaging.

Conclusions
In this work, a core-shell nanoparticle of NaGdF 4 @CaCO 3 -PEG was designed as an activatable MR/US dual-modal imaging contrast for cancer diagnosis, which is triggered by the acidic environment. The general idea behind this OFF/ON responsive MR imaging behavior consists of quenching the sphere Gd 3+ relaxation effects by coating NaGdF 4 with a layer of hydrophobic CaCO 3 to limit water availability. At acidic aqueous solution, CaCO 3 was dissolved to generate CO 2 bubbles which is used to obtain US signal. Meanwhile, a strong MRI enhancement can be activated upon dissolution of CaCO 3 and release of the previously silenced NaGdF 4 into the aqueous solution. In vivo results demonstrated the strong dual-modal magnetic resonance/ ultrasonic imaging abilities of NaGdF 4 @CaCO 3 -PEG at the tumor site with an acidic environment. We expect that this work may provide a new insight for strategies to design nanomaterials with responsive dual-modal imaging abilities.

Methods
Materials. Gd(CH 3 CO 2 ) 3 and PEG 8000 were purchased from Sigma Aldrich Co., Ltd. Acetone and cyclohexane were obtained from Sinopharm Chemical Reagent Co., Ltd. 1-Octadecene and Methanol was purchased from Alladin Company. Oleic acid was purchased from Alfa Aesar. Cell Counting Kit (CCK-8) was obtained from Dojindo laboratories. Penicillin-streptomycin, fetal bovine serum (FBS), and Dulbecco's Modified Eagle Medium (DMEM) were purchased from Gibco BRL. All of these materials were used as received without further purification. Ultrapure water was used throughout.   34,38 . In brief, 24 mL of Gd(CH 3 CO 2 ) 3 (0.2 M) was added to a 500 mL of two-necked round-bottomed flask containing 36 mL of oleic acid. When the mixture was heated to 150 °C for 30 min, 84 mL of 1-octadecene was added and then heated to 150 °C to form the Gd-oleate precursor solution. Then, the reaction system was cooled down to 60 °C naturally and a mixture of 8 mL NaOH (1 M) and 16 mL NH 4 F (0.4 M) in methanol was added and stirred for 30 min. The resultant solution was heated at 110 °C for 30 min to remove the methanol, oxygen and water under argon atmosphere. Thereafter, the solution was heated to 300 °C and kept for 1.5 h before cooling down to room temperature. The as-prepared nanoparticles were precipitated by additional of ethanol, collected by centrifugation at 6000 rpm for 3 min, and washed with ethanol for several times. The NaGdF 4 nanoparticles are re-dispersed in 50 mL of cyclohexane for further application.
Synthesis of the core-shell NaGdF 4 @CaCO 3 -PEG nanoparticles. The core-shell NaGdF 4 @CaCO 3 nanoparticles were prepared through a facile microemulsion method with using NaGdF 4 nanoparticles as templates 39 . Briefly, 5 mL NaGdF 4 nanoparticles, 10 mL cyclohexane, 3.45 mL Triton X-100, and 3.2 mL 1-hexanol were added to a flask and mixed thoroughly, followed by the addition of a 30 mM solution of calcium chloride (800 μL) to form a well-dispersed water-in-oil emulsion. Then, the solution of sodium carbonate (40 μL, 2.92 M) was added. After keeping moderate stirring for 8 h of the mixture, the as-prepared NaGdF 4 @CaCO 3 nanoparticles were collected by centrifugation and then re-dispersed in 10 mL of ultrapure water. Thereafter, 2 mL PEG 8000 (0.2 M) aqueous solution was added and stirred for another 8 h, then the as-prepared nanoparticles were collected by centrifugation at 10000 rpm for 15 min, and re-dispersed in 20 mL of ultrapure water for further application.
Characterization of the NaGdF 4 @CaCO 3 -PEG nanoparticles. The morphology and the size of the obtained NaGdF 4 @CaCO 3 -PEG nanoparticles were performed on a Tecnai F20 transmission electron microscope (TEM, FEI Company, Hillsboro, OR) with an accelerating voltage of 200 kV 40 . For the TEM experiment, the suspension of nanoparticles was dropped onto a carbon-coated copper grid, followed by drying naturally. Scanning transmission electron microscopy (STEM) was obtained using a JEM-2010 electron microscope (JEOL, Japan) to characterize the chemical composition of NaGdF 4 @CaCO 3 -PEG nanoparticles 40 . Dynamic light scattering (DLS) experiments were recorded at 25 °C on a NanoZS (Malvern Instruments, UK) with a detection angle of 173°, and a 3 mW He-Ne laser operating at a wavelength of 633 nm. FT-IR spectra were performed on a Fourier transform infrared spectrometer (Perkin-Elmer, Spectrum-2000) over the spectral region of 4000 cm −1 to 400 cm −1 41 .

Statistical analysis.
All quantitative data were expressed as the mean ± standard deviation (SD). Graph Pad Prism version 6.0 was used for statistics analysis. Statistical analysis among different groups was performed using student T-Test. The P < 0.05 was considered as statistically significant.
Animals. All mice (4-5 weeks old; weighing: 18-22 g) were obtained from the Center for Animal Experiment of Fujian Medical University (License No: SCXKmin2012-0002), and housed at constant temperature (22 ± 2 °C) and 60% relative humidity, with a light/dark (hours) cycles of 12/12. All animal procedures were conducted in accordance with the approved guidelines of Animal Ethics Committee of Fujian Medical University.
Cytotoxicity assay. The LN3 and NIH3T3 were cultured in Dulbecco's Modified Eagle Medium (DMEM) containing 10% fetal bovine serum and 1% penicillin streptomycin at 37 °C in a humidified atmosphere (5% CO 2 ). The in vitro cytotoxicity was investigated by the Cell Counting Kit (cck-8) assay. In detail, LN3 and NIH3T3 cells were cultured in a 96-well cell-culture plate at a density of 10 4 (100 μL) cells per well for 24 or 48 h, respectively. Then, the medium was replaced with 100 μL fresh medium containing various concentrations of NPs. After 24 or 48 h incubation, the medium was removed. Then, 100 μL of fresh medium and 10 μL of CCK-8 were added and incubated for another 2 h. The absorbance was measured by a Bio-Rad Model-680 microplate reader Visualization of CO 2 Bubble Generation from NaGdF 4 @CaCO 3 -PEG nanoparticles. To observe the bubble generation characteristic of NaGdF 4 @CaCO 3 -PEG, the aqueous dispersion was dropped on a glass slide, followed by drying naturally. Then, PBS buffer with various pH values from 5.0 to 7.4 was dropped on the samples, and the CO 2 bubble image from NaGdF 4 @CaCO 3 -PEG was obtained by an optical microscope at room temperature 10, 11 . In Vitro US Imaging at Various pH. In vitro US imaging of NaGdF 4 @CaCO 3 -PEG was performed in phosphate buffer solutions at various pH conditions (7.4, 7.0, 6.8, and 5.0) 41 . An optically transparent phantom gel plate, which was made by embedding a 500 μL Eppendorf tube in the agarose gel (3%, w/v) and then removing the tube after the phantom gel had cooled, was used for the in vitro experiments. Aqueous nanoparticle solutions (10 mg•mL −1 ) were prepared at various pH. US images were obtained using Vevo 2100 imaging system operated at 21 MHz of a static state using a contrast mode. The change of US intensity for each sample was measured up to 180 min, and the US intensity of the water as control was subtracted from the sample intensity for the normalization.
In Vitro MR Imaging at Various pH. In vitro MR imaging of NaGdF 4 @CaCO 3 -PEG was performed in phosphate buffer solutions at various pH conditions (7.4, 7.0, 6.8, and 5.0) using T 1 -weighted MRI on a 0.5 T NMI20-Analyst NMR system (Niumag Corporation, Shanghai, China) to evaluate the contrast-enhancement effect 42 . Relaxivity and MRI phantom studies at 0.5 T magnetic field. A series of NaGdF 4 @CaCO 3 -PEG nanoparticle aqueous solutions with different Gd concentrations (12.5, 10, 7.5, 5.0, 2.5, and 1.25 mM) were prepared for MRI phantom and relaxivity studies. All experiments were performed on a 0.5 T NMI20-Analyst NMR system (Niumag Corporation, Shanghai, China) [42][43][44] . The longitudinal relaxation times (T 1 ) were measured using an inversion recovery (IR) sequence. The longitudinal (r1) was determined from the slope of the plot of 1/T 1 against the Gd concentration (mM).

In vivo US imaging of Xenograft Tumor.
To form a solid tumor in nude mice, the LN3 xenograft tumor was estabilished in 4-week-old male nude mice by injecting 10 6 LN3 cells into the right thigh of mice. After injection, tumor-bearing nude mice were kept for 10~14 days to achieve a tumor size around 80 mm 3 . Then, 200 μL of normal saline (NS) containing NaGdF 4 @CaCO 3 -PEG (2 mg•mL −1 ) was injected by an intratumoral injection. After injection, the tumor was imaged with the Vevo 2100 imaging system.