Fe₃O₄ nanoparticles were first synthesized via a hydrothermal co-precipitation approach (Ding et al., 2016). HA was wrapped onto the surface of nanoparticles through electrostatic interactions (Figure 1). XRD was used to investigate the patterns of nanoparticles. Both bare Fe₃O₄ (black) and Fe₃O₄@HA (red) nanoparticles showed same peaks in (220), (311), (400), (422), (511), and (440) planes (Figure 2A), indicating that the incorporation of HA into the system happened in a physical absorption way rather than a chemical reaction. FTIR results (Figure 2B) showed the existence of characteristic signals of HA in Fe₃O₄@HA nanoparticles, suggesting successful functionalization of HA corona. Furthermore, TGA measurements were used to quantitatively identify the content of HA binding on the surface of Fe₃O₄@HA nanoparticles. Compared with bare Fe₃O₄ nanoparticles, the modified Fe₃O₄@HA nanoparticles exhibited an increase in weight loss from 12.75 to 32.39% (Figure 2C). The content of HA on the surface of nanoparticles is 19.64%. These results together indicate that we have successfully synthesized Fe₃O₄@HA nanoparticles with an HA loading of around 20% by weight.

We next characterized the morphology and size of the resulting Fe₃O₄@HA nanoparticles. TEM imaging clearly revealed spherical nanoparticles with a uniform size distribution (Figure 3A). The size of the nanoparticles was further measured by ImageJ and estimated to be around 10.25 ± 2.11 nm (Figure 3B). Since we did not stain the nanoparticles, the contrast of the HA corona should be relatively low and even negligible. Therefore, the size we measured here is possibly the size of bare Fe₃O₄ itself. The actual size of Fe₃O₄@HA should be considered larger than the size here with the extra HA corona. In addition, we examined the hydrodynamic size and physical stability of Fe₃O₄@HA in PBS buffer via DLS. The results showed that the hydrodynamic size of Fe₃O₄@HA nanoparticles is in between 270–310 nm (Figure 4), which is much larger than the size measured by TEM. This is plausibly caused by slight aggregation or interaction among nanoparticles in the solution, as also observed by others in similar systems (Li et al., 2014). The intensity averaged particle size measured by DLS often reflects a bit more on large-sized particles even though the majority are small ones (Fischer and Schmidt, 2016). With TEM imaging, we were able to identify the single nanoparticles. Monitoring the hydrodynamic size of Fe₃O₄@HA for 14 days (Figure 4) showed that the system did not form a super large aggregation or precipitate out within this period of time. Therefore, the bare Fe₃O₄ nanoparticles were successfully stabilized by coated HA. The zeta potential of Fe₃O₄@HA was measured as 32.3 ± 0.8 mV.

Since magnetic Fe₃O₄ nanoparticles can shorten the transverse T ₂ relaxation time of water protons so as to enhance the contrast, we next characterized the T ₂ of protons in the Fe₃O₄@HA solutions of various Fe concentrations. The results (Figure 5) clearly showed that with the increase in the Fe concentration, Fe₃O₄@HA nanoparticles were able to cause a decrease in the magnetic resonance (MR) intensity in the T ₂-weighted mode. The transverse relaxivity r ₂ (the transverse relaxation rate per mM of iron) was further calculated by linear fitting of the relaxation rate 1/T ₂ vs. Fe concentration (Figure 5). The r ₂ of the Fe₃O₄@HA nanoparticles was estimated at around 314 mM⁻¹s⁻¹. This number is higher than the numbers reported in the literature (Shi et al., 2008; Cai et al., 2012; Li J. et al., 2013; Li et al., 2014), which could be possibly caused by the thinner layer of HA corona due to electrostatic interactions compared with the chemical conjugated ones. Water molecules can penetrate easily through the thinner HA corona and interact with Fe₃O₄ in the core of the nanoparticles. These results suggested that Fe₃O₄@HA nanoparticles with such high r ₂ can be used as a potential candidate for T ₂-weighted MRI.

Given that all the contrast agents are eventually used for in vivo detection and diagnosis, they should be extremely biocompatible with the body. We therefore investigated the hemocompatibility of the Fe₃O₄@HA nanoparticles, which is one of the crucial prerequisites before the in vivo studies. The hemolysis assay was performed at different Fe concentrations (0.025, 0.05, 0.1, 0.2, 0.4, 0.8, and 1.6 mM) with water and PBS as controls (Figure 6) for 2 h incubation at 37°C. The absorbents were collected after incubating Fe₃O₄@HA nanoparticles with RBC suspension (Figure 6). Along with PBS solution, we did not notice the strong hemolysis effect in the samples with Fe₃O₄@HA, while the water control showed an obvious hemolysis phenomenon. By comparing the absorbance at 541 nm, we found that the percentage of hemolysis is almost negligible (less than 3%) under the current Fe concentrations. Therefore, we concluded that our Fe₃O₄@HA nanoparticles were hemocompatible, warranting further investigation in vivo.

In order to function as imaging agents for living cells, the nanoparticles have to get into cells and also be compatible with cells. We therefore evaluated the cytotoxicity and cellular uptake of Fe₃O₄@HA nanoparticles. The nanoparticles at different Fe concentrations (0.025–0.4 mM) were incubated with HeLa cells for 24 h at 37°C. The morphologies of HeLa cells treated with Fe₃O₄@HA were observed by phase contrast microscopy. The images (Figure 7A) showed that the morphologies of the nanoparticle groups were similar to those treated with PBS, and no obvious changes were observed, indicating that Fe₃O₄@HA nanoparticles are compatible with cells under the current concentration range. This observation was further verified by in vitro cytotoxicity studies. The viabilities of HeLa cells were measured after incubation with Fe₃O₄@HA nanoparticles (0.025, 0.5, 0.1, 0.2, and 0.4 mM) for 24 h by MTT assay. Clearly, the Fe₃O₄@HA nanoparticles did not exert any cytotoxicity until the Fe concentration was at 0.2 mM (Figure 7B). Further increase in the Fe concentration to 0.4 mM slightly decreased the cell viability by less than 10% compared with the cells treated with PBS. These results, together with the observation of cell morphology, demonstrated that Fe₃O₄@HA nanoparticles are biocompatible and almost non-cytotoxic at the Fe concentration up to 0.4 mM.

Since HA has high binding affinity with the CD44 receptor that is overexpressed in HeLa cells, we next assessed the targeting ability of HA toward the model HeLa cell line. Cells were treated with PBS and Fe₃O₄@HA nanoparticles at concentrations of 0.025, 0.5, 0.1, 0.2, and 0.4 mM. After 4 h of incubation, the cells were washed with PBS three times, digested, centrifuged, and resuspended in PBS for cell counting. Cells were further lysed, and the Fe uptake was quantified by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). The negative control group was created by pretreating HeLa cells with HA (10 μM in DMEM) for 2 h to saturate the CD44 receptor and then incubating with Fe₃O₄@HA nanoparticles. As shown in Figure 7C, the HeLa cells showed a dose-dependent cellular uptake of Fe₃O₄@HA. The amount of nanoparticles entered the cells increased with increasing Fe concentration in both the Fe₃O₄@HA group and the negative control HA-blocked group. More importantly, the normal Fe₃O₄@HA group showed higher cellular uptake compared with the HA-blocked group at all concentrations. Quantitively, the amounts of Fe uptake in the Fe₃O₄@HA group were 1.7, 2.6, 1.5, 1.6, and 1.9 folds compared with those in the HA-blocked group at the incubation concentrations of 0.025, 0.5, 0.1, 0.2, and 0.4 mM, respectively (Figure 7C). These observations suggested that the HA corona on the surface of Fe₃O₄@HA nanoparticles can mediate the specific interactions with the CD44 receptor expressed by HeLa cells and realize an active targeting mechanism. The results also perfectly elaborate on the design principle of the work.

To verify the feasibility of using Fe₃O₄@HA nanoparticles as contrast agents, we next performed an in vitro MRI study. HeLa cells were co-incubated with Fe₃O₄@HA nanoparticles at various Fe concentrations (0.025, 0.5, 0.1, 0.2, and 0.4 mM) for 6 h at 37°C with PBS as a control. Furthermore, T ₂-weighted MRI of the HeLa cells was performed and analyzed. As shown in Figure 8A, the MR images of the cells treated with Fe₃O₄@HA presented lower signal intensity than the PBS control group. The signal also exhibited a dose-dependent behavior, which decreased with the increase in the Fe concentration (Figure 8B). These results indicated the effective role of Fe₃O₄@HA nanoparticles in providing adequate imaging contrast, and the dose-dependent decrease in signal intensity is probably a result of the concentration-dependent cellular uptake.

After demonstrating the possibility of in vitro MRI in cells, we further investigated the use of Fe₃O₄@HA nanoparticles for in vivo MRI in a HeLa xenografted tumor model. When the tumor reached a size of around 150 mm³, Fe₃O₄@HA nanoparticles were administered intravenously at Fe concentration of 10 mM in 100 μL PBS. MRI was then performed at 1, 2, 4 and 8 h post-injection. In the group of blocked HA, it (100 μM in 50 μL PBS) was intratumorally injected to saturate the CD44 receptor 1 h before administration of nanoparticles. The T ₂-weighted MR images of Fe₃O₄@HA and HA-blocked groups are shown in Figures 9A,B, respectively. We found that the MR signals in both groups decreased after injection. Based on more quantitative assessment (Figure 9C), we found that the highest contrast enhancement happened at 1 h post injection and then gradually recovered. In the Fe₃O₄@HA group, the signals decreased by 66, 50, 30, and 20% at 1, 2, 4, and 8 h post-injection, respectively, compared with that before injection. However, in the HA-blocked group, the signals only decreased by 29% at 1 h post injection and had already recovered back to similar intensity as that before injection at 4 h. Direct comparison between Fe₃O₄@HA and HA-blocked groups showed significant differences, indicating that Fe₃O₄@HA nanoparticles achieved more effective tumor targeting in the HA active state than in the HA-blocked state. The slight decrease in the MR intensity in the HA-blocked group could be a result of passive targeting because of the EPR effect, and the recovery of the intensity could be mainly attributed to the metabolism process that cleared nanoparticles out of the body, but the significantly lower MR signals in the Fe₃O₄@HA group than in the HA blocked group at the same time point clearly indicated the role of the HA-mediated specific tumor targeting pathway in addition to the EPR effect. These results are consistent with our in vitro cellular uptake studies, which again confirmed that Fe₃O₄@HA nanoparticles could target the tumor through an active targeting mechanism, realizing more effective MRI both in vitro and in vivo.

Hyaluronic acid (6 KDa) was purchased from Zhenjiang Dong Yuan Biotechnology Corporation (Zhenjiang, China). Iron (III) chloride hexahydrate (FeCl₃.6H₂O), iron (II) chloride tetrahydrate (FeCl₂.4H₂O), and sodium hydroxide were purchased from Aladdin Ltd. (Shanghai, China). 3-(4,5Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was supplied by Thermo Fisher Scientific, Ltd. (Waltham, MA). Fetal bovine serum (FBS), Dulbecco’s modified Eagle’s medium (DMEM), penicillin-streptomycin, and trypsin were obtained from Gibco Life Technologies Co. (Grand Island, NY). Regenerated cellulose dialysis membranes with a molecular weight cut-off (MWCO) of 50 KDa were acquired from Fisher (Pittsburgh, PA).

The nanoparticles were synthesized via a facile controlled co-precipitation method. Briefly, 30 mg of HA was first dissolved in 25 ml Milli-Q water (resistivity higher than 18.2 MΩ cm, a Milli-Q Plus 185 water purification system (Millipore, Bedford, MA)) and stirred in a three-neck round-bottom flask to obtain a homogeneous solution. The solution was degassed with bubbling N₂ to remove O₂ and heated to 80°C. Then, 0.36 g of FeCl₃.6H₂O and 0.132 g of FeCl₂.4H₂O were dissolved in 7.5 ml O₂-free Milli-Q water and added to the HA solution. After stirring for 15 min under N₂ protection, 10 ml of O₂-free Milli-Q water containing 1 g of NaOH was quickly added to the mixture under mechanical stirring of 1200 rpm. The reaction was continued for another 2 h at 80°C. After cooling down to room temperature, the black solution was 1) centrifuged at a low speed of 600 rpm to get rid of the big aggregations and 2) magnetically separated and re-dispersed in the Milli-Q water. The obtained solution was further dialyzed (molecular weight cut-off at 50 KDa) against Milli-Q water for 2 days to remove free ions and HA. A small quantity of the Fe₃O₄@HA solution was subjected to freeze-drying, and the leftover was stored at 4°C.

The crystalline structure of the Fe₃O₄ and Fe₃O₄@HA nanoparticles was characterized by X-ray diffraction (XRD) in a 2θ range of 20–80°, using a D/max 2550 PC X-ray diffractometer (Japan, Rigaku Cop.) with Cu Kα radiation (λ = 0.154,056 nm) at 40 kV and 200 mA. Fourier transform infrared (FTIR) spectra of Fe₃O₄ and Fe₃O₄@HA nanoparticles were obtained by using a Nexus 670 spectrometer (Thermo Nicolet Corporation, Madison, WI). Thermogravimetric analysis (TGA) was performed in a temperature range of 30–900°C with a heating rate of 20°C/min under nitrogen using a TG209 F1 (NETZSCH Instruments Co., Ltd., Germany) thermogravimetric analyzer.

The TEM samples were prepared by adding 10 μL of Fe₃O₄@HA nanoparticle (Fe concentration 0.1 mM) solution onto a carbon-film-coated copper grid (400 square mesh, Electron Microscopy Sciences, Hatfield, PA, United States), and the excess solution was wicked with a filter paper. The grid was air-dried before imaging and was then imaged using a FEI Tecnai 12 TWIN transmission electron microscope (100 kV). A SIS Megaview III wide-angle CCD camera was used to acquire the TEM images. The samples for dynamic light scattering measurements were prepared at a Fe concentration of 0.2 mM, and then, the hydrodynamic size of Fe₃O₄@HA nanoparticles was measured by a Malvern Zetasizer Nano ZS model ZEN3600 (Worcestershire, U.K.) equipped with a standard 633-nm laser. Three repeated measurements for each sample were determined to give the average values and standard deviations.

The Fe concentration of Fe₃O₄@HA nanoparticles was previously determined using the Leeman Prodigy Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) system (Hudson, NH03051). Sample solutions at Fe concentrations of 0.00156, 0.00313, 0.00625, 0.0125, 0.025, 0.05, and 0.1 mM were prepared by dilution of stock solution before measurements. T ₂ relaxometry was performed by using a 0.5-T NMI20-Analyst NMR Analyzing and Imaging system (Shanghai Niumag Corporation, China). Instrumental parameters are as follows: a point resolution of 156 mm × 156 mm, section thickness of 0.6 mm, TR of 4000 ms, TE of 60 ms, and number of excitation of 1. The T ₂ relaxivity was determined by linear fitting of 1/T ₂ vs. Fe concentration.

Mouse blood (1.5 ml) collected from the inner canthus vein plexus was mixed with 3.5 ml of PBS. Pure red blood cells (RBCs) were obtained via repeated centrifugation/redispersion processes (2000 rpm, 10 min, three times). The RBCs were then diluted with 5 ml of PBS for further use. A measure of 100 μL of the obtained RBC suspension was mixed with 900 μL PBS (negative control), water (positive control), and Fe₃O₄@HA nanoparticle solution in PBS at various Fe concentrations (0.025, 0.05, 0.1, 0.2, 0.4, 0.8, and 1.6 mM). After 2 h incubation at 37°C, sample solutions were centrifuged at 10,000 rpm for 15 min. The absorbance of the supernatant for each sample at 540 nm was then measured via a Lambda 25 UV-Vis spectrophotometer (PerkinElmer, Boston, MA). The hemolysis rate was calculated as follows: hemolysis rate ( % ) = ( A ( sample,  540  nm ) - A ( negative,  540  nm ) ) / ( A - ( positive,  540  nm ) A ( negative,  540  nm ) ) × 100 % .

HeLa cells were purchased from the Institute of Biochemistry and Cell Biology (the Chinese Academy of Sciences, Shanghai, China) and cultured in DMEM containing 5% FBS and 1% antibiotics at 37°C and 5% CO₂.

HeLa cells (10,000 cells/well) were seeded in 96-well plates overnight. The Fe₃O₄@HA nanoparticles at various Fe concentrations (0.025, 0.05, 0.1, 0.2, and 0.4 mM) were next incubated with the HeLa cells for another 24 h in 200 μL of DMEM. HeLa cells treated with PBS were used as a control. The morphology of HeLa cells was further observed by phase contrast microscopy (Leica DM IL LED inverted phase contrast microscope) at a magnification of 200 times.

In vitro cytotoxicity was further quantitively confirmed by the MTT assay. Similar to the protocols described earlier, HeLa cells (10,000 cells/well) were seeded in 96-well plates overnight. The Fe₃O₄@HA nanoparticles at various Fe concentrations (0.025, 0.05, 0.1, 0.2, and 0.4 mM) were then incubated with the HeLa cells for another 24 h in 200 μL DMEM. Furthermore, the cells were rinsed three times with PBS and then incubated in 100 μL of FBS-free DMEM medium containing 10% MTT for 4 h. After removal of the medium, the MTT assay was performed according to the manufacturer’s instructions. For each concentration, three parallel wells were measured to give the average values and standard deviations.

HeLa cells (2 × 10⁶ cells/well) were seeded in 6-well plates for overnight adherence. The seeded plates were divided into two groups, and the cell medium was replaced on the second day. One group was replaced with fresh DMEM medium, and the other group was replaced with fresh DMEM containing HA (10 μM) that pre-saturates the overexpressed CD44 receptor in HeLa cells. After incubation for another 2 h, the medium was removed, and the cells were washed with PBS three times. Then, the cells were further incubated with fresh DMEM containing Fe₃O₄@HA nanoparticles at various Fe concentrations (0.025, 0.05, 0.1, 0.2, and 0.4 mM) for 4 h at 37°C. Next, the cells were washed three times with PBS, digested by trypsinization, centrifuged (1000 rpm, 5 min), and resuspended in PBS for cell counting. The remaining cells were centrifuged (1000 rpm, 5 min) and lysed using 0.5 ml aqua regia solution (nitric acid/hydrochloric acid, v/v = 1:3) for one day. Finally, we diluted the samples with PBS, and the cellular uptake of the Fe₃O₄@HA nanoparticles was evaluated by inductively coupled plasma-atomic emission spectroscopy (ICP-AES).

The animal study protocol was approved by the Ethics Committee of Scientific Research and Clinical Trials of the First Affiliated Hospital of Zhengzhou University. HeLa cells (5 × 10⁶ cells/flask) were seeded into 25 cm² culture flasks with 5 ml DMEM overnight at 37°C. On the second day, Fe₃O₄@HA nanoparticles at different Fe concentrations (0.025, 0.05, 0.1, 0.2, and 0.4 mM) in 5 ml fresh medium were replaced. After 6 h incubation, the HeLa cells were rinsed three times with PBS, digested by trypsinization, centrifuged (1000 rpm, 5 min), and resuspended in 1 ml (containing 0.5% agarose) PBS in a 2-ml eppendorf tube. The T ₂ -weighted MRI of HeLa cells was performed on a 3.0 T Signa HDxt superconductor clinical MR system (GE Medical Systems, Fairfield, CT). 2D spin-echo MR images were obtained with the parameters of 2 mm slice thickness, TR/TE 2000/96.2 ms, FOV 6 × 6 cm, and 256 × 256 matrix.

For the experiment, 6-week-old female BALB/c nude mice were purchased from SPF (Beijing) Biotechnology Co., Ltd. HeLa cells (2 × 10⁶ in 100 ml PBS) were subcutaneously injected into the left back of the mice. After 3–4 weeks, when the tumor reached a size ∼150 mm³, the mice were randomly divided into two groups (n = 3 per group). One group was intratumorally injected with HA (100 μM) in 50 μL PBS 1 h before the injection of Fe₃O₄@HA contrast agents. Then, both groups of mice were anesthetized by an intraperitoneal injection of pentobarbital sodium (40 mg/kg). The Fe₃O₄@HA nanoparticles (Fe concentration = 10 mM, in 100 μL PBS) were then intravenously injected into mice. The in vivo tumor MRI studies were conducted at different time points (1, 2, 4, and 8 h post injection) using a 3.0 T Signa HDxt superconducting clinical MR system attached with a custom-built animal receiver coil. 2D spin-echo MR images were obtained with the parameters of 2 mm slice thickness, TR/TE 2000/96.2 ms, FOV 6 × 6 cm, and 256 × 256 matrix. The T ₂ -weighted MR images before administration were also obtained as controls.