Hyaluronic acid-modified manganese-chelated dendrimer-entrapped gold nanoparticles for the targeted CT/MR dual-mode imaging of hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is the most common malignant tumor of the liver. The early and effective diagnosis has always been desired. Herein, we present the preparation and characterization of hyaluronic acid (HA)-modified, multifunctional nanoparticles (NPs) targeting CD44 receptor-expressing cancer cells for computed tomography (CT)/magnetic resonance (MR) dual-mode imaging. We first modified amine-terminated generation 5 poly(amidoamine) dendrimers (G5.NH2) with an Mn chelator, 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA), fluorescein isothiocyanate (FI), and HA. Then, gold nanoparticles (AuNPs) were entrapped within the above raw product, denoted as G5.NH2-FI-DOTA-HA. The designed multifunctional NPs were formed after further Mn chelation and purification and were denoted as {(Au0)100G5.NH2-FI-DOTA(Mn)-HA}. These NPs were characterized via several different techniques. We found that the {(Au0)100G5.NH2-FI-DOTA(Mn)-HA} NPs exhibited good water dispersibility, stability under different conditions, and cytocompatibility within a given concentration range. Because both AuNPs and Mn were present in the product, {(Au0)100G5.NH2-FI-DOTA(Mn)-HA} displayed a high X-ray attenuation intensity and favorable r1 relaxivity, which are advantageous properties for targeted CT/MR dual-mode imaging. This approach was used to image HCC cells in vitro and orthotopically transplanted HCC tumors in a unique in vivo model through the CD44 receptor-mediated endocytosis pathway. This work introduces a novel strategy for preparing multifunctional NPs via dendrimer nanotechnology.


Synthesis of {(Au
The procedure used to synthesize {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} was similar to what has been previously reported, with slight modifications 26,27  Extensive dialysis was applied to remove the excess reactants and by-products in the reaction mixture, and subsequent lyophilization was used to obtain the final product.
The small positive surface charge of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs allowed the commonly used acetylation step to be omitted. In the control group, the CD44 receptors were blocked with free HA at a concentration of 25 mM for half an hour before the targeted NPs were injected via the tail vein.
Characterization techniques. DLS and zeta potential measurements were conducted using a Malvern Zetasizer Nano ZS model ZEN3600 (Worcestershire, U.K.) with a standard 633-nm laser. Prior to taking the measurements, the samples were dissolved in water (0.1 mg/mL). We used TEM (JEOL 2010F, Tokyo, Japan) with an accelerating voltage of 200 kV to characterize the morphology of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs. TEM samples were prepared by depositing a dilute NP suspension (6 μ L) onto carbon-coated copper grids and allowing it to air-dry before observation. The TEM image selections were random, and the average size and size distribution of more than 300 NPs were recorded using ImageJ software. ICP-AES (Leeman Prodigy, USA) was employed to analyze the composition of Au and Mn within the multifunctional NPs. A GE LightSpeed VCT imaging system (GE Medical Systems) was used for CT scanning at 80 mA, 100 kV, and a slice thickness Cytotoxicity assay and cell morphology observations. An MTT assay was applied to measure the in vitro cytotoxicity of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs in a routine culture environment with 10% FBS (heat-inactivated, 1% penicillin-streptomycin) and 5% CO 2 at 37 °C.
HCCLM3 cells suspended in medium were seeded in a 96-well plate at a density of 1 × 10 4 cells/well with 200 μ L per well and were incubated overnight. NPs with an Mn concentration ranging from 0-100 μ g/mL were added into each well, and the cells were incubated for an additional 24 h. The mixture was then carefully removed, and the cells were washed twice with phosphate-buffered saline (PBS). Then, 20 μ L of MTT solution (5 mg/mL in PBS) was added into each well, and the cells were cultured for another 4 h at 37 °C and 5% CO 2 . To dissolve the insoluble formazan crystals, the medium was carefully discarded and replaced with 200 μ L of DMSO. Finally, the absorbance of each sample was measured at 570 nm using a Thermo Fisher Scientific Multiskan MK3 ELISA reader (Thermo Fisher Scientific, Hudson, NH).
The morphology of the HCCLM3 cells was observed to evaluate the cytotoxicity of the {(Au 0 ) 100 G5. In vitro cellular uptake assay. The cellular uptake of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs in vitro was assessed by flow cytometry. HCCLM3 cells 51 , which highly express CD44, were used for this experiment. Cells were seeded into each well of a 12-well plate at a density of 2 × 10 5 cells/well and incubated with 2 mL of DMEM overnight. NPs with Mn concentrations of 10, 25, 75, and 100 μ g/mL were added into the wells, and the cells were incubated for an additional 4 h. Cells treated with PBS and cells treated with HA followed by NPs were used as controls.
In vivo CT/MR imaging. All  Statistical analysis. The experimental data were analyzed by a single-factor analysis (one-way ANOVA).
Our NMR results ( Figure S1a), which were obtained using a previously reported method 27 , indicated that each G5 dendrimer was linked with approximately 23.6 DOTA moieties. These linked DOTA moieties can be used to chelate Mn for T 1 MR imaging. Based on the feeding ratio, the theoretical number of DOTA moieties linked to each G5 dendrimer was 30, but the actual number was slightly smaller. We utilized the same method to calculate that approximately 13.6 HA and 3.7 FI moieties were linked to each G5 dendrimer ( Figure S1b,c).
The zeta potential and hydrodynamic size of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs were recorded ( Table 1). The final product clearly exhibited a slight positive charge (+ 6.1 mV) due to the HA coupled to the dendrimer terminal amines and the omission of acetylating the remaining dendrimer terminal amines 26,27 . The small amount of positive charge on the NP surface may have been caused by some of the G5 amines being used to stabilize the AuNPs. DLS was used to assess the hydrodynamic size of the developed NPs dissolved in water ( Table 1). The {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs clearly displayed a hydrodynamic size of 245.3 nm. The NPs exhibited both an acceptable PDI and excellent colloidal stability.
The morphology and size of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs were observed by TEM (Fig. 2). The formed NPs were nearly spherical in shape (Fig. 2a) and had a mean diameter of 2.1 nm with a relatively uniform size distribution (Fig. 2b). The size measured by DLS was larger than that measured by TEM, likely because DLS characterizes the size of {(Au 0 ) 100 G5.  Stability. The stability of these NPs is essential for their biomedical application. To evaluate the stability of the prepared {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs, the NPs were dissolved in different solutions (i.e., water, PBS, or cell culture medium). After the NPs were stored for one month at room temperature, no precipitate was observed ( Figure S2). Meanwhile, as described in the literature 28 , we also measured the hydrodynamic size of the NPs after 7 days of storage at room temperature. The hydrodynamic diameter of the NPs was 266.5 nm, which was not significantly different from the value observed before storage.  (Fig. 3b).   52 , Au is superior to iodine in terms of X-ray attenuation (e.g., Omnipaque) because Au has a higher atomic number. Therefore, AuNPs have been widely applied in CT contrast agents. The results clearly show that the CT value increases as the Au concentration increases (Fig. 3c,d). By linearly fitting the attenuation intensity versus the Au concentration, a dose-dependent relation was obtained.
In vitro cytotoxicity. Prior to biomedical application, the cytocompatibility of the developed {(Au 0 ) 100 G5.
NH 2 -FI-DOTA(Mn)-HA} NPs was evaluated using an MTT colorimetric assay (Fig. 4). After 24 h of incubation with HCCLM3 cells, no significant cytotoxicity from the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs was observed at any of the tested Mn concentrations. Compared to the PBS control, no significant difference (P > 0.05) in HCCLM3 cell viability was observed even at the highest Mn concentration (100 μ g/mL). Cell viability remained at more than 80%. Clearly, the produced NPs are cytocompatible in the given Mn concentration range.
To further evaluate NP cytotoxicity, after the HCCLM3 cells were treated with {(Au 0 ) 100 G5. NH 2 -FI-DOTA(Mn)-HA} NPs at Mn concentrations of 10, 20, 50, 75, and 100 μ g/mL for 24 h, the morphology of the cells was then observed by phase-contrast microscopy ( Figure S3). The results showed that the morphology of the HCCLM3 cells treated with the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs at Mn concentrations of 10-100 μ g/mL ( Figure S3b-f) was similar to that of the cells in the control group (treated with PBS) ( Figure S3a). The MTT  Flow cytometry assay. The NPs were modified by FI molecules and could be analyzed by flow cytometry through the binding of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs to the target cells. Cells with blocked CD44 receptors were used as a control. At different Mn concentrations, the fluorescence intensity of HCCLM3 cells with unblocked CD44 receptors was much stronger than that of cells with blocked CD44 receptors ( Figures  S4 and 5). The enhanced cellular uptake of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs should be related to the modified HA molecules that can specifically target HCCLM3 cells via the CD44 receptor-mediated pathway.   and retention (EPR) effect as well as active targeting. However, in the control group, this effect is likely only attributable to NP accumulation via the EPR effect because of the blocked CD44 receptors. Based on the quantitative changes in the CT and MR signal intensity values over time after the actively targeted NPs were injected, the tumor CT density and MR signal intensity values are highest at 0.5 h post-injection and partly recover at 1 h post-injection. The CT and MR signal intensity values of the tumors treated with the NPs in the experimental group were obviously higher than those of the free HA in the control group at the same time point (P < 0.001). At 1 h post-injection with free HA + NPs, the NPs began to be metabolized and leave the tumor site, which led to the recovery of the CT value and MR signal, precisely the opposite of the significantly higher CT density and MR signal intensity values maintained after the NP injection without free HA. The CT/MR imaging results of the orthotopic tumors showed that the synthesized {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs can be used as a nanoprobe for effective, targeted CT/MR dual-mode imaging in vivo. Our results suggest that {(Au 0 ) 100 G5. NH 2 -FI-DOTA(Mn)-HA} NPs have the ability to target cancer cells through an HA-mediated targeting pathway because the targeting ability was weakened when CD44 receptors were blocked. Meanwhile, we also measured the CT and MR signal intensity values of normal livers, and no significant differences were found between the two groups at the different time points (Fig. 8). These results also clearly showed that the CT density and MR signal intensity values were greater in the tumor tissue than in the normal liver tissue. The CT/MR images of orthotopic HCC in vivo could be used to easily distinguish tumor tissue from normal tissue.  In vivo biodistribution. It is crucial to know the biodistribution of the synthesized {(Au 0 ) 100 G5. NH 2 -FI-DOTA(Mn)-HA} NPs prior to their application in advanced in vivo biomedical imaging. The in vivo biodistribution of the {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs in major organs, such as the liver, spleen, kidneys, lungs, and heart, as well as in the tumor, was analyzed by ICP-AES at 24 h post-injection (Fig. 8). After the injection of both NP samples, the Au concentration in the measured organs was obviously higher in the treatment group than in the blank control group. The main uptake of Au occurred in the liver, spleen and lungs in the treatment group, while relatively little uptake occurred in the other tissues (i.e., heart, kidneys, and tumor). The biodistribution data of the Au indicated that the particles could be delivered to the tumor tissue after escaping the reticuloendothelial system. The targeting property of these {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs was clearly demonstrated, as the Au concentration in the tumor region of the mice treated with the targeted {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs was double that in mice treated with the non-targeted NPs. HA plays an important role in this process by enabling the highly efficient delivery of NPs to the tumor region in vivo for targeted CT/MR tumor imaging.

Conclusions
In summary, {(Au 0 ) 100 G5.NH 2 -FI-DOTA(Mn)-HA} NPs with X-ray attenuation favorable for CT imaging and r 1 relaxivity suitable for T 1 -weighted MR imaging were developed and applied in an orthotopic HCC tumor model. Dendrimers were used as a template to entrap AuNPs within and chelate Mn ions onto the template surface. EDC coupling chemistry was used to couple HA molecules onto the G5 dendrimer surfaces, thereby endowing the NPs with the ability to actively target CD44 receptor-expressing cancer cells. The favorable characteristics of these multifunctional NPs, including water solubility, colloidal stability, and biocompatibility, make them extremely attractive for potential use in the CT/MR imaging of tumors in vivo via an active, HA-mediated targeting pathway. With the application of this strategy, various dual-or multimode imaging contrast agents may be designed for the accurate diagnosis of various types of malignant tumors.