In Vivo Demonstration of Enhancing Gas-filled Microbubble Magnetic Susceptibility with Iron Oxide Nanoparticles

INTRODUCTION Gas-filled microbubbles have been recently employed in therapeutic applications due to their unique cavitation and sonoporation properties. Local microbubble cavitation by spatially focused ultrasound can be applied in achieving site-specific release of incorporated drugs or genes inside microbubbles. Microbubbles can potentially be visualized by MRI and serve as an intravascular susceptibility contrast agent in vivo due to the induction of large local magnetic susceptibility difference by the gas-liquid interface. However, microbubble susceptibility effect is relatively weak when compared with other intravascular MR susceptibility contrast agents. Previous theoretical and phantom studies have indicated that, by embedding or coating magnetic nanoparticles, the magnetic susceptibility of the shell can be increased, thus enhancing the microbubble susceptibility effect. In this study, we aim to further demonstrate the synergistic effect of gas core with iron oxide nanoparticles in achieving the overall microbubble susceptibility effect and characterize in vivo enhancements of microbubble susceptibility effects by entrapping iron oxide nanoparticles using dynamic susceptibility weighted MRI at 7 T. METHODS All MRI experiments were performed on a 7 T Bruker MRI scanner. Synthesis of iron oxide nanoparticles entrapped polymeric microbubbles: Polymeric microbubbles (PMBs) with monocrystalline iron oxide nanoparticles (MIONs; MGH) incorporated or entrapped in shells were produced as previously described (Fig. 1). To demonstrate the synergistic effect of the gas core with MIONs, MION-entrapped solid microspheres (without gas core) were also prepared with the same procedures without performing lyophilization for comparison. Characterization of PMBs: ΔR2 was measured by acquiring multi-echo GE signals continuously without phase encoding for 2 min as previously described. The absolute iron contents were determined with inductively coupled plasma mass spectrometry (ICP-MS) for MION-entrapped PMBs and solid microspheres. Their r2 relaxivities were computed as the R2 in the initial uniform suspended state normalized with their respective iron contents. Animal Preparation and Microbubbles Administration: Normal SD rats (~250-350 g) (N = 6) were injected intravenously with 0.2 mL of microbubble suspension (~3.5% volume fraction) at a rate of 0.6 mL/min to avoid possible microbubble destruction due to high pressure and shear stress under femoral vein catheterization. The susceptibility effect of microbubbles was compared with SonoVue® (Bracco), by single dose of ~3.5% volume fraction injection using identical injection protocol and imaging sequence. MRI: Anatomical images were acquired using 2D FLASH sequence with resolution = 0.2 × 0.2 × 2 mm. Dynamic susceptibility weighted liver MRI was performed with respiratory-gated single-shot GE-EPI sequence using TR ≈ 1000 ms, TE = 10 ms, FA = 90o, FOV = 50 × 50 mm, slice thickness = 2 mm, acquisition matrix = 64 × 64, BW = 221 kHz and NEX = 1. Image Analysis: GE-EPI images were first co-registered using AIR5.2.5. ΔR2 maps were computed as previously described. To quantify the ΔR2 values, ROIs were manually drawn in homogeneous liver region (LV) and the region covering blood vessel (BV) based on the high resolution FLASH images. Two-tail paired Student’s t test was employed to compare the liver ΔR2 induced by MION-free and MION-entrapped PMBs with p < 0.05 considered as statistically significant. Assuming that ΔR2 is proportional to microbubble concentration C(t) at time t, C(t) were approximately modeled with a gamma-variate function by curve fitting given the relatively long injection time and the limited lifetime of microbubbles in vivo. Full width at half maximum (FWHM) and time-to-peak were then measured from the fitted C(t) time courses. RESULTS AND DISCUSSIONS Table 1 shows the measurements of the ∆R2 values and relaxivities for MION-entrapped PMBs (N = 6), MION-free PMBs (N = 6) and MION-entrapped microspheres (N = 2). ∆R2 for MION-entrapped PMBs was significantly higher than that of MION-free PMBs and MION-entrapped solid microspheres. The relaxivity results underscore the importance of both gas core and entrapped MIONs in achieving strong overall microbubble susceptibility effect, as suggested by the earlier theoretical study. Figure 2(a) illustrates one of the preinjection GE-EPI T2-weighted images, while (b) shows the corresponding postinjection GE-EPI T2-weighted image with the maximum susceptibility contrast for MION-entrapped PMB injection. Figure 3 depicts the typical T2-weighted signal time courses during PMBs injection with ROIs indicated in Fig. 2(a). Similar signal time course was also observed during MION-free PMB and SonoVue® injection. For MION-entrapped PMBs, the T2-weighted signals after microbubble injection did not return to the preinjection baseline, this may be caused by accumulation of released MION in Kupffer cells after in vivo microbubble clearance. The computed ∆R2 maps for MION-free PMBs and MION-entrapped PMBs in the same animal were shown in Fig. 4, reflecting the blood distribution in tissue vasculature. Table 2 shows the in vivo measurements of the ∆R2, FWHM and time-to-peak of MION-free PMBs, MION-entrapped PMBs and SonoVue® in liver among all rats studied. ∆R2 induced by MION-entrapped PMBs was found to be significantly higher than that by MION-free PMBs (p < 0.05), demonstrating that in vivo microbubble susceptibility can be enhanced by the MIONs entrapped in shells, as suggested by the earlier theoretical and phantom studies.