Cation Exchange Strategy to Construct a Targeting Nanoprobe for Enhanced Positive MR Imaging Capability


 Background: Excellent imaging performance and good biocompatibility of contrast agents are considered as prerequisites for accurate tumor diagnosis and treatment. Results: Herein, a novel imaging nanoprobe with actively targeting performance based on ultrasmall paramagnetic iron oxide (USPIO) was constructed by a facile cation exchange strategy followed by conjugation with transferrin (Tf). The stable gadolinium (Gd3+) chelation endows the nanoparticles (NPs) with a low value of r2/r1 (1.28) and relatively high r1 value of 3.2 mM-1s-1, enabling their use in T1-weighted positive MR imaging.Conclusion: This constructed transferrin modified gadolinium-iron chelate nanoprobe, named as TUG, shows high biocompatibility within a given dose range. More importantly, compared with clinically used Gd-based small molecule contrast agents, the obtained TUG can be more engulfed by breast cancer cells, showing much enhanced T1-weighted positive MR imaging in either subcutaneous or in situ tumor models of breast cancer. This novel nanoprobe holds enormous promise to be utilized as a targeting contrast agent with high efficacy for T1-weighted positive MR imaging.


Background
Magnetic resonance imaging (MRI) has become one of the most common clinical examination methods due to its good tomographic capabilities and high contrast of soft tissue [1][2][3][4] . But the sensitivity and e ciency of cancer diagnoses have generally been impeded by the dilemma in identifying tumor tissue from normal tissue [5][6][7] . To overcome the above obstacles, gadolinium (Gd)-based small molecule chelates are currently used in clinical practices as contrast agents to improve the contrast [8][9][10][11] . However, these gadolinium-based T 1 contrast agents contain free toxic gadolinium, which potentially leads to renal brosis and other toxicity concerns, and these agents still have some other shortcomings, such as short blood circulation time, no target speci city, etc 12,13 . To reduce the side effects of gadolinium-based contrast agents, an iron-based contrast agent is becoming a promising alternative. Although those contrast agents with iron oxide nanoparticles (NPs) as the cores have higher biocompatibility compared to gadolinium-based contrast agents 14,15 , most iron-based contrast agents are negative contrast agents that lead to insurmountable disadvantages, such as di culty in distinguishing from bleeding, metal deposits, or calci cation [16][17][18] . Therefore, the construction of a positive contrast agent based on iron species to reduce the content of gadolinium ion and to improve the biocompatibility of the contrast agent is highly demanded 19 .
With the decrease of the size of iron oxide NPs, the speci c surface area of NPs increases and the energy exchange of T 1 becomes faster 20,21 . The tissue signal collected from T 1 -weighted image increases and T 2 effect decreases, which can be used for T 1 -weighted MRI 17,22,23 When the size of ultra-small paramagnetic iron oxide (USPIO) NPs is less than 5 nm, the performance of negative contrast agents will be greatly weakened, while the performance of positive contrast agents will be greatly improved 24,25 .
However, practically the T 1 -weighted MR imaging performance of USPIO is not super to that of clinically used gadolinium-based small molecule contrast agents 26,27 due to seven unpaired electrons of gadolinium. Herein, to improve the T 1 -weighted MRI performance of USPIO, a facile cation exchange strategy was adopted to solvate the carboxyl groups covering ultra-small paramagnetic iron oxide NPs prepared by a solvothermal method with Gd(III) ions. Compared with the gadolinium-based contrast agent, Gd(III) ion-induced self-assembled NPs, named as USPIO@Gd(III) (Abbr. UG), present good biocompatibility and stability with tunable gadolinium content.
It is well known that small-molecule contrast agents currently used in clinical practice do not have the tumor-targeting ability. To increase the accumulation of contrast agents in tumor site, we use chemical crosslink to attach transferrin on the surface of UG. Due to the high expression of transferrin receptors in tumor cell membranes 28 , the binding of transferrin to receptors can increase the number of nanoparticles entering the tumor cells. The prepared UG was thus further modi ed with transferrin to form the targeted imaging nanoprobe, namely TUG, which shows good colloidal stability, high r 1 relaxivity, and good biocompatibility. More importantly, TUG shows enhanced cellular uptake e cacy by breast cancer cells, indicating good tumor-targeting imaging and longer MRI window time performance.

Preparation and Characterization of TUG
A one-step solvothermal method is proposed to prepare ultrasmall paramagnetic iron oxide with stable surface modi cation of sodium citrate (Fig. 1). Due to a large amount of sodium citrate on the surface of nanoparticles, USPIO has a strong charge repulsion force, which is bene cial to maintain excellent colloidal stability in different liquid environments 29,30 . The UG NPs were obtained by the cation exchange between Gd(III) and sodium citrate on the surface of USPIO. Subsequently, carboxyl groups covered the nanoparticles and amino groups on the transferrin were bonded via an EDC/NHS reaction 31,32 , resulting in the nal product TUG.
The photograph of TEM reveals that the prepared ultrasmall iron oxide nanoparticles are uniformly dispersed with a stable diameter of about 2 nm (Fig. 2a). After addition of Gd ions and Tf, the obtained TUG formed by the assembly of ultrasmall iron oxide shows good dispersion with a diameter of about 50 nm (Fig. 2b). SEM element line scanning analysis of TUG clearly shows three elements of iron, gadolinium, and oxygen homogeneously co-present in the obtained nanoprobe (Fig. S1). Figure 2c shows the zeta potentials and hydrodynamic sizes of the USPIO, UG, and TUG. In detail, the zeta potential of USPIO is -40.0 mV due to the abundant carboxyl groups, while after addition of Gd(III) and Tf, the zeta potential increased to -20.0 mV and − 18.0 mV, respectively, indicating successful surface modi cation (Fig. 2c). The hydrodynamic sizes of USPIO, UG, and TUG are 2 nm, 50 nm, and 82 nm, respectively, which also implying the formation of nanoprobe with successful surface modi cation of transferrin.
Furthermore, the results of zeta potentials and hydrodynamic sizes of TUG in different media (H 2 O, PBS, and RPMI) for 7 days indicate that TUG can maintain colloidal stability for a long time, which is suitable to in vivo application (Fig. S2). During the synthesis process, the feed ratio of Fe/Gd was set to 10:1, and the actual ratio of Fe/Gd in TUG measured by ICP-OES was 16.7:1. Additionally, ICP-OES results show that the molar ratio of Fe/Na in USPIO nanoparticles and TUG is 3.7 and 61.3, respectively, con rming that sodium ions have been successfully exchanged by Gd ions (Fig. 2d). The conjugation of transferrin was con rmed by FTIR and TGA ( Fig. 2e and Fig. 2f). After modi cation with transferrin, the absorption peaks at 3390 cm − 1 , 2930 cm − 1 , 1652 cm − 1 , and 584 cm − 1 correspond to the O-H, C-H, N-H, and Fe-O stretching of TUG, indicating the successful modi cation of Tf. The TGA curve (Fig. 2f) shows a modest decline at the beginning of the test implying a 10% weight loss before 100 o C caused by the adsorbed water in TUG. The total weight loss from 200 o C to 500 o C due to the thermal decomposition of organics is around 55%, in which the amount of Tf can be calculated to be 37.5%. Additionally, the stability experiment of TUG veri es that less than 5% free gadolinium ions released from TUG after 8 hours in the saline (Fig. 2g), which ensures its hypotoxicity.

T 1 Relaxometry and MRI Imaging
The r 2 and r 1 values of USPIO and TUG can be calculated by plotting the reciprocal of relaxation time (1/T 2 or 1/T 1 ) as a function of Fe concentration ( Fig. 2h and Fig. 2i). It is found that, for USPIO, the r 1 and r 2 values are 0.1 mM − 1 s − 1 and 0.6 mM − 1 s − 1 , respectively, with the r 2 /r 1 ratio of 6.0. Interesting, for TUG, the r 1 and r 2 values are 3.2 mM − 1 s − 1 and 4.1 mM − 1 s − 1 , respectively, with the r 2 /r 1 ratio of 1.3. Compared with USPIO, TUG shows higher r 1 and lower ratio of r 2 /r 1 , which could be ascribed to the shortened longitudinal relaxation time due to the synergistic effect of the obtained cluster structure.

Phagocytosis, Cytotoxicity and in vitro MRI of TUG in 4T1 cells
The cytocompatibility of the contrast agents is the premise of applying TUG in vivo. As shown in Fig. 3a, after 24 h co-culture, TUG shows negligible cytotoxicity to 4T1 cells (the cell viability of 4T1 cells is above 90% within the given concentration from 0-100 µg/mL). It is worth mentioning that when the coincubation time prolonged to 48 h under iron content of 100 µg/mL, the cell survival rate remained above 80%, indicating good cytocompatibility of prepared TUG. Furthermore, we validated TUG MRI performance at the cellular level before performing in vivo animal imaging. As shown in Fig. 3b-c, regardless of TUG group or commercial contrast agent Magnevist (Gd-DTPA), the MR signal of 4T1 cells after co-incubation increases with increasing concentration of different contrast agents. Interestingly, at the same concentration, the MR signal of the TUG group is remarkably higher than that of the commercial contrast agent group, indicating that TUG has better imaging performance than the commercial contrast agent. Besides, the introduction of transferrin increases the cell phagocytosis of nanoparticles due to the overexpressed transferrin receptors of 4T1 breast cancer cells.
We thus evaluated the difference in phagocytosis of TUG and UG by ICP-OES and Prussian blue staining.
As shown in Fig. 3e, at the same dose, the phagocytosis of TUG by 4T1 cells reaches 36.1 μg per 100,000 cells, while the phagocytosis of non-targeted nanoparticles is only 8.3 μg per 100,000 cells. The amount of nanoparticles consumed by cells in TUG group was 4.3 times of the UG group and the statistical results proved that the modi cation of transferrin considerably increases the phagocytosis of cells by nanoclusters. The results of Prussian blue staining show that 4T1 cells treated with TUG show clear blue color, while almost no blue signals appeared in the saline-combined UG group, which was consistent with the ICP results ( Fig. 3d-e).

MRI of TUG Nanoprobes in vivo
The enhancement effect of TUG as an MRI contrast agent in vitro was studied by comparing and analyzing T 1 -weighted MRI images (Fig. S3). The MRI of TUG shows that T 1 signal increases with the increase of the Gd proportion. Then, MR images of mice were obtained for further study of the enhancement effect of MRI ( Fig. 4 and Fig. S6). T 1 -weighted MR images of whole tumor areas were acquired before and 1 h and 8 h after injection by recording the tumor area signal values from all periods and drawing average luminance histogram of MR image tumor area. In the subcutaneous tumor model, the signal peak was detected in mice injected with Magnevist 10min after injection, since Magnevist only passively accumulates in the peripheral vascular tissues of tumor sites by EPR effect, resulting in a low concentration of TUG in the tumor region, and almost unchanged relaxation time of regional hydrogen proton T 1 . Eight hours after the injection of TUG, the MR images of the tumor areas of the mice are not only brighter than those before the injection, but also brighter than those during the peak in the tumor areas of the mice injected with Magnevist. In the tumor models in situ, Magnevist's signal peak time is about 30 min, and the magnitude of Magnevist's signal peak occurs in less than 0.5 hours after the injection of TUG. It is proved that the T 1 -weighted imaging effect of TUG is superior to Magnevist in tumor regions. Notably, the peak time varies in different tumor models, which could be ascribed to the fact that the blood supply of subcutaneous tumor is not as rich as that of breast tumors in situ, leading to longer peak time.

Safety Evaluation of TUG
The long-term biosafety of TUG in living mice after intravenous injection was also evaluated. Section pictures of organs stained by Hematoxylin and eosin (H&E) showed that physiological morphologies of organs including heart, lung, liver, spleen, and kidney of Balb/c female mice remained normal for 14 days after injection of TUG even at a high dose of 15 mg/kg (Fig. 5). To determine whether TUG hurts liver function and renal function, the levels of enzymes and chemicals in serum such as alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN) and creatinine (CRE) were measured and no abnormals found. Also, the hematocrit, hemoglobin, mean corpuscular volume, mean corpuscular hemoglobin, lymph cells, platelet, red blood cells, red cell distribution, and white blood cells remained almost unchanged after injection of TUG. All the above results indicate high biosafety of the prepared TUG.

Synthesis of TUG Nanoprobe
Sodium citrate stabilized USPIO was prepared according to our previous reported literature 33 . Then, under magnetic stirring at 0 o C the gadolinium chloride aqueous solution (5.5 mL, 2 mg/mL) was dropwise added into USPIO aqueous solution (10 mL 5 mg/mL). After 2 hours of reaction, the obtained USPIO@Gd(III) NPs (Abbr. UG) were dialyzed against deionized water (6 times, 2 L) for 2 days in a dialysis membrane (MWCO 8,000 Da). Subsequently, Tf was combined with UG as a whole through the reaction between -NH 2 and -COOH in the presence of EDC/NHS. According to the instructions of the reagents, 1 mL EDC (13 µmol) and 1 mL NHS (13 µmol) were added into 5.6 mL of UG (ice bath precooling) under magnetic stirring for 1.5 hours. After that, 2.4 mL of Tf (17 mg/mL) was dropwise added into UG/EDC/NHS mixtures. After 5 h magnetic agitation reaction at room temperature, TUG nanoprobe was puri ed by membrane dialysis (MWCO 8,000 Da) to remove the unreacted free EDC and NHS. The nal products were named as TUG.

Characterization of TUG
The morphology of USPIO and UG were characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Cation exchange between sodium and gadolinium ions was characterized by Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES). Size distribution, zeta potentials, Fourier transform infrared (FTIR), thermogravimetry (TGA), and MR imaging were conducted to characterize the nanoclusters.

Colloidal Stability and Gd(III) Release Test
The colloidal stability of nanoparticles is a crucial index to evaluate the possibility of their subsequent biological applications. We estimated the colloidal stability of the prepared nanoparticles by monitoring the hydrated particle size changes of TUG dispersed in deionized water, PBS, and serum-containing medium for 7 days. The release of Gd(III) from chelating agents is one of the potential toxic and adverse effects of clinically used Gd(III)-based contrast agents. Therefore, we tested the cumulative release of Gd(III) from the dialysis bag to determine the stability of the nanocluster prepared by the cation exchange strategy. Brie y, 3 mL TUG solution (Fe concentration 10 g/mL) was added into a dialysis bag (MWCO 8,000 Da), and the dialysis bag was then placed in a 50 mL centrifuge tube containing 20 mL of saline and shaken in a constant temperature shaker (37 o C, 120 rpm). 5 mL of external liquid in the tube was collected at a settled time (15 min, 30 min, 1 h, 2 h, 4 h, and 8 h) and 5 mL of new saline was replenished. All samples were then tested by ICP-OES for Gd(III) measurement.

Toxicity Analysis and Cellular Uptake
CCK8 method was selected to study the cytotoxicity of TUG to 4T1 cells (mouse breast cancer cells). 4T1 cells were seeded at a cell density of 1 × 10 4 per well in 96-well plate containing 0.2 mL of RPMI medium, and incubated at 37 o C in a humidi ed atmosphere with 5% CO 2 for 12 h. Next, the nutrient-depleted medium was replaced by fresh medium containing TUG of different concentrations at 37 o C, 5% CO 2 for 12 h and 24 h. And then the culture medium was replaced by fresh serum-free medium (90 µL) containing CCK8 (10 µL, 10%) for 4 h incubation. Absorbance at 450 nm in each well was measured by Bio-Tek ELX800 spectrophotometric microplate reader (Vermont, America), based on which cell viability was calculated. All experiments were conducted three times in parallel. The phagocytosis of UG and TUG by 4T1 cells was evaluated by using ICP-OES and Prussian blue staining. Brie y, 4T1 cells were seeded in a 6-well plate at a density of 5 × 10 5 cells/well for 16 h Then the culture medium was replaced by fresh medium containing UG or TUG in each well for 4 h incubation. The cells were washed three times with PBS before the Prussian blue staining and ICP-OES test.
3.6 In vitro T 1 -weighted MR Imaging Performance of TUG 4T1 cells were cultured in the same environment as mentioned above for 8 h and then the culture medium was replaced by a fresh medium with or without TUG, or UG (C Gd 0.1-0.00125 mM). After the subsequent 2 h treatment, the cells were washed three times by PBS, treated with trypsin for 3 min, and then collected by centrifuge at 1000 rpm for 5 min. The obtained cells were digested with Aqua regia for Gd measurement by ICP-OES. For MRI, the obtained cells were dispersed in 0.5 mL of agarose (0.5%) for MR scanning using a 3.0 T clinical MR imaging system under the following parameters: FOV = 60 × 60 mm, matrix = 256 × 256, section thickness = 2 mm, TR = 1200 ms, and TE = 17 ms.

In vivo MR Imaging of Breast Tumor
Animal experiments were carried out according to protocols approved by the Southern Medical University Committee Animal Care and Use Committee, and also following the policy of the National Ministry of Health. 4 weeks old Balb/c female nude mice (body weight ≈ 18 g, n = 12) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd, and randomly divided into two groups with six in each group. 3 × 10 6 4T1 cells/mice were implanted under the skin of the right-back and mammary gland to establish a subcutaneous tumor model and in situ tumor model of breast cancer. When the breast tumor nodules reached a volume of about 200 mm 3 after approximately 7 days, each group (n = 6) was divided into two subgroups (n = 3) and injected with an equal dose of Magenevist and TUG (1 mg/mL, 250 µL) through the tail vein for later analysis of the signal enhancement in the tumor site. After injection, the mice were anesthetized by intraperitoneal injection of 5% chloral hydrate (120 µL). The T 1 -weighted images were obtained by a 0.5 T MRI system at 0 h, 2 h, 4 h, 6 h, 8 h. The parameters of MRI in vivo were set as follows: TR = 360 ms, TE = 13 ms, matrix = 256 × 160, section thickness = 2 mm, and FOV = 80 × 80 mm. The brightness of the tumor was measured by MATLAB 7.0. The enhancement of brightness showed an increased T 1 -weighted signal.

In vivo Biodistribution and Toxicity Assessment
After the nanoprobes were delivered to the mice through the tail vein, their toxicity to the mice determines the fate of biomedical applications. The in vivo toxicity of the nanoprobe was analyzed 14 days after injection of different doses of the nanoclusters, and the blood biochemistry, changes in blood routine indexes, and tissue staining of organs were analyzed.

Statistical Analysis
The signi cant difference between the experimental statistics is analyzed by One-way ANOVA and Tukey's multiple comparison tests. When the P-value < 0.05, statistics were regarded to be signi cantly different. P < 0.05 was indicated by (*), P < 0.01 by (**) and P < 0.001 by (***).

Conclusions
In summary, a facile cation exchange strategy followed by conjugation with transferrin (Tf) was developed to construct a novel TUG nanoprobe with excellent biocompatibility. The low value of r 2 /r 1 (1.28) and relatively high r 1 value of 3.2 mM − 1 s − 1 enable the obtained TUG an excellent candidate as T 1weighted MR imaging contrast agent. The in vivo results of either the subcutaneous or in situ tumor models of breast cancer con rm that the TUG has the obviously enhanced T 1 -weighted MR imaging performance to the clinically used Gd-based small molecule contrast agents. Therein, TUG with transferrin modi cation enabling high speci c binding performance to the transferrin receptor of breast cancer also greatly contributes to the high capability of T 1 -weighted MR imaging. Such targeted speci c binding nanoprobes are expected to be highly promising in biomedical applications, such as e cient MRI probe for cancer diagnosis.

Con icts of interest
The authors declare that they have no competing interests.

Availability of data and materials
All data generated or analyzed during this study are included in this published article (and its additional le).

Funding
This work was supported by the National Natural Science Foundation of China (Grant No. 51772316, 81630046 and 81471659) Figure 1 Schematic illustration of the preparation of TUG nanoprobes.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.