Tumor-targeted Gd-doped mesoporous Fe3O4 nanoparticles for T1/T2 MR imaging guided synergistic cancer therapy

Abstract In this study, a novel intelligent nanoplatform to integrate multiple imaging and therapeutic functions for targeted cancer theranostics. The nanoplatform, DOX@Gd-MFe3O4 NPs, was constructed Gd-doped mesoporous Fe3O4 nanoparticles following with the doxorubicin (DOX) loading in the mesopores of the NPs. The DOX@Gd-MFe3O4 NPs exhibited good properties in colloidal dispersity, photothermal conversion, NIR triggered drug release, and high T1/T2 relaxicity rate (r1=9.64 mM−1s−1, r2= 177.71 mM−1s−1). Benefiting from the high MR contrast, DOX@Gd-MFe3O4 NPs enabled simultaneous T1/T2 dual-modal MR imagining on 4T1 bearing mice in vivo and the MR contrast effect was further strengthened by external magnetic field. In addition, the DOX@Gd-MFe3O4 NPs revealed the strongest inhibition to the growth of 4T1 in vitro and in vivo under NIR irradiation and guidance of external magnetic field. Moreover, biosafety was also validated by in vitro and in vivo tests. Thus, the prepared DOX@Gd-MFe3O4 NPs would provide a promising intelligent nanoplatform for dual-modal MR imagining guided synergistic therapy in cancer theranostics.


Introduction
Despite the advance of modern medical technology, cancer remains as a great threat to human health (Torre et al., 2016). Clinically, the conventional therapies, such as surgery, radiotherapy, and chemotherapy still cannot effectively conquer this disease ascribed to their low therapeutic efficacy and severe adverse effects (Vanneman & Dranoff, 2012;Gotwals et al., 2017;Lee et al., 2018). Fortunately, nanoparticle-based strategies have emerged in the development of alternative cancer therapies owing to their advantages for overcoming the limitations associated with the conventional treatments (Perez-Herrero & Fernandez-Medarde, 2015). More recently, multifunctional nanoparticles have attracted more attention owing to their versatility that can integrate multiple functions such as molecular imaging, chemotherapeutics, active tumor targeting, phototherapy, and immunotherapy, into one nanoplatform. These intelligent nanomaterials would provide a novel strategy in nanomedicine for cancer theranostics (Arranja et al., 2017;Liu et al., 2018).
Among these versatile functions, medical imaging occupies the primary role in precise cancer diagnosis, which is also able to guide the cancer treatment process (Kijima et al., 2014). Up to date, various imaging methods have been developed for biomedical imaging, including fluorescence imaging (FL), magnetic resonance imaging (MRI), computed tomography imaging (CT), position emission tomography (PET), photoacoustic imaging (PA), etc. (Cuevas & Shibata, 2009;Berges et al., 2010;Fan et al., 2014;Tempany et al., 2015). Among the various medical imaging techniques, magnetic resonance imaging (MRI) is a very powerful and noninvasive imaging tool to provide high 3D spatially resolved images with the information on the anatomy, function and metabolism of tissues in vivo (Fernando et al., 2013;Lu et al., 2013). Normally, MR imaging is performed in T 1 or T 2 mode based on the T 1 -weighted contrast agents (CAs) or T 2weighted CAs, respectively, to improve the sensitivity of MR imaging. However, each contrast agent possesses its own merits and limitations in MR imaging applications (Czeyda-Pommersheim et al., 2017).
Clinically, Gd 3þ -based paramagnetic complexes, such as Gd-DTPA, are widely used in T 1 -weighted MR imaging to obtain brighter images. The T 1 -weighted MR enhancement is realized by reducing the longitudinal relaxation times to provide positive contrast (Courant et al., 2012;Gupta et al., 2015;Phukan et al., 2018). However, these Gd 3þ complexes are always prepared into small molecules, which would lead to fast clearance from body and hamper the further engineering of the complexes (Boehm & Heverhagen, 2018). On the other side, superparamagnetic iron oxide nanoparticles (SPIONs) are extensively investigated as T 2 Cas (Qin et al., 2015). These SPIONs are able to reduce transverse relaxations times to produce darker images and negative contrast in the T 2 -weighted MR imaging practice (Cheng et al., 2011(Cheng et al., , 2012. Unfortunately, for the low signal body regions, it would be difficult to distinguish the damaged tissues with hemorrhages, calcification, fat, blood clots and other possible artifacts, which would hamper the further potential in MR imaging. Thus, the development of dual-modal T 1 -T 2 CAs would provide superior contrast effect in both T 1 and T 2weighted MR images to conquer problems associated with the CAs with signal modality Zhou et al., 2017;. Recently, photothermal therapy (PTT) induced by near infrared (NIR) laser has become a promising therapeutic strategy to destroy tumor in a noninvasive way (Hou et al., 2018;Liu et al., 2019). Normally, the photothermal tumor ablation can be achieved by using the laser absorbing agents to convert laser energy into thermal energy locally at tumor site. Till date, a variety of nanomaterials have been developed as laser absorbing agents, such as some gold nanoparticles, carbon-based nanomaterials, and semiconductor nanostructures (Robinson et al., 2010;You et al., 2012;Huang et al., 2017). Particularly, besides the MRI practice as T 2 CAs, the iron oxide nanoparticles (Fe 3 O 4 NPs) also exhibit good photothermal converting ability for PTT in cancer. For example, Shen et al. successfully prepared individual magnetic Fe 3 O 4 NPs and applied them in photothermal therapy with NIR irradiation (Shen et al., 2015). Zhou et al. also proposed iron/iron oxide core/shell nanoparticles for magnetic targeting MRI and near-infrared photothermal therapy (Zhou et al., 2014).
Herein, as shown in Scheme 1, we propose a novel multifunctional nanoplatform for magnetic targeted T 1 -T 2 dual MR imaging guided combined cancer therapy. The nanoplatform (DOX@Gd-MFe 3 O 4 NPs) was constructed by Gd-doped mesoporous Fe 3 O 4 nanoparticles following with the doxorubicin (DOX) loading in the mesopores of the NPs. The prepared DOX@Gd-MFe 3 O 4 NPs exhibited good colloidal dispersity, superior magnetization, high r 1 and r 2 relaxavity, excellent photothermal conversion, and NIR-triggered drug release. The in vitro cellular uptake was investigated by confocal laser scanning microscope and Prussian blue staining, which the influence of NIR laser and external magnetic field was also explored. Benefiting from the outcome of photothermal conversion and cellular uptake, the in vitro combined PTT and chemotherapy was evaluated by live/dead cell staining and MTT assay. Next, we moved to investigate the magnetic targeted T 1 /T 2 -weighted MR imaging on 4T1 tumor bearing mice in vivo. Furthermore, the synergistic PTT and chemotherapy under the guidance of MRI was also verified on 4T1 tumor bearing mice in vivo. Finally, the longterm biosafety of the Gd-MFe 3 O 4 NPs was explored in vitro and in vivo. Thus, we expect the prepared DOX@Gd-MFe 3 O 4 NPs would favor tumor targeted drug delivery for T 1 /T 2 dual modal MR imaging guided combined tumor therapy. NPs (3 mg) were dispersed in 3 mL DI water, and then EDC (30 mg) and DOTA-NHS (3 mL, 12 mM) were added under stirring for 6 h. Thereafter, the product was washed thrice with DI water to remove excess EDC and DOTA-NHS. Next, DOTA-MFe 3 O 4 NPs (5 mg) were dispersed in 5 mL GdCl3 solution (12 mM) and maintained for 12 h for the Gd chelation. Finally, the product was centrifuged and washed to obtained the Gd-MFe 3 O 4 NPs. In addition, the content of Gd and Fe element in the Gd-MFe 3 O 4 NPs were measured by an inductively coupled plasma-mass spectrometry (ICP-MS).

DOX loading in Gd-MFe 3 O 4 NPs
Free DOX (3 mg) was firstly dissolved in 1 mL PBS buffer and mixed with above Gd-MFe 3 O 4 NPs (2 mg/mL, 5 mL). Then, the resultant mixture was stirred at 500 rpm overnight in the dark and then centrifuged at 10,000 rpm for 10 min. Afterwards, the NPs were washed thrice with DI water to remove the unloaded DOX and the supernatant solution was collected to investigate the loading efficacy. Finally, the DOX@Gd-MFe 3 O 4 NPs were centrifuged, washed, and freezedried for the further use. The drug loading efficiency (LE) and encapsulation efficiency (EE) can be calculated according to following equations: W initial is the weight of DOX in initial solution, W remanent is the weight of DOX in supernatant, and Gd-MFe 3 O 4 is the weight of Gd-MFe 3 O 4 NPs added in solution.

Characterization of DOX@Gd-Fe 3 O 4 NPs
The shape and morphology of DOX@Gd-Fe 3 O 4 NPs was studied by scanning electron microscopy (SEM, SU8010, Hitachi, Tokyo, Japan) and transmission electron microscopy (TEM, FEI Tecnai G2 Spirit). The particle size distribution and surface zeta potentials of DOX@Gd-Fe 3 O 4 NPs were investigated by a Zetasizer Nano ZS90 (Malvern Instruments Ltd, Malvern, UK). The chemical structure of DOX@Gd-Fe 3 O 4 NPs was explored using a Fourier Transform Infrared spectrometer (FTIR,Bruker Tensor27,Billerica,MA). N 2 adsorptiondesorption isotherms were assessed using a surface area and pore characterization system (ASAP 2020, AutoChem II). UV-vis-NIR absorption spectra was determined by an UV-vis-NIR spectrometer. X-ray diffraction (XRD) measurements was carried out by a high-resolution X-ray diffractometer (XRD, X' Pert PRO Multi-Purpose, XRD). The magnetization was evaluated by vibrating sample magnetometer (VSM, Lake Shore Cryotronics 7404, Westerville, OH) at 300 K.

Photothermal conversion test of DOX@Gd-MFe 3 O 4 NPs
The photothermal conversion test was explored on 0.5 mL DOX@Gd-MFe 3 O 4 NPs suspension under 808 nm laser irradiation (FC-808-10W, Changchun New Industries Optoelectronics Technology Co. Ltd, Changchun, China). The DOX@Gd-MFe 3 O 4 NPs were suspended in PBS buffer at various concentrations (PBS, 0.05, 0,10, 0.25, 0.50, and 1.00 mg/ mL) and placed in a quartz cuvette. Then, the suspensions were irradiated by NIR laser at different intensities (0.2, 0.5, 1, 1.5, 2 W/cm 2 ) for 5 min. An IR thermal imaging camera (FLIR E60; FLIR Systems, Inc., Wilsonville, OR) was deployed to monitor the temperature elevation of the solutions. Furthermore, we investigated the photothermal stability by repeatedly exposing the DOX@Gd-MFe 3 O 4 NPs suspension to NIR laser (laser on) and then naturally cooled to room temperature (laser off) for 4 cycles.

In vitro DOX release profile from DOX@Gd-MFe 3 O 4 NPs
The in vitro release of DOX from DOX@Gd-MFe 3 O 4 NPs was performed using the standard dialysis method in the presence and absence of NIR laser. Briefly, DOX@Gd-MFe 3 O 4 NPs were suspended in 2 mL PBS buffer (containing 2 mg DOX), transferred in a dialysis bag (MWCO: 3500 Da), and then placed in a tube containing 50 mL PBS buffer. Thereafter, 1 mL samples were withdrawn at pre-determined time points and replaced by 1 mL fresh PBS buffer. Simultaneously, NIR laser (2 W/cm 2 ) was used to trigger the release at 4 and 12 h time point. The concentration of released DOX in each sample was analyzed by a microplate reader.
2.6. In vitro cell studies 2.6.1. Cell culture 4T1 mouse breast cancer cells (4T1) and NIH3T3 mouse embryonic fibroblast cells (HSF) were obtained from American Type Culture Collection (ATCC, Manassas, VA) and maintained in RPMI medium containing 10% (v/v) FBS and 1% 1% penicillin-streptomycin. The cells were incubated at 37 C in a humidified 5% CO 2 atmosphere.

In vitro cell uptake study
The in vitro cell uptake of the DOX@Gd-MFe 3 O 4 NPs was first studied by the confocal laser scanning microscopy (CLSM). Briefly, 4T1 cells were seeded on sterile coverslips in a 12well plate (3 Â 10 4 cells/well) for 24 h to allow cell attachment. Then, DOX@Gd-MFe 3 O 4 NPs (DOX¼ 5 lg/mL, NPs ¼ 48 lg/mL) were added in each well and co-incubated with cells for 2 h in the presence and absence of a magnet (magnetic field: 0.05 T). Simultaneously, NIR laser (1.5 W/cm 2 , 5 min) was used to irradiate the cells during the NPs incubation to explore the effect of NIR laser to the intracellular drug release. Thereafter, the cells were washed thrice with PBS solution and fixed with 4% PFA. Then, the cells were mounted with gold anti-fade mounting medium containing DAPI and observed with a CLSM (CLSM, LSM880, Zeiss, Jena, Germany). The cellular uptake of DOX@Gd-MFe 3 O 4 NPs was further performed by the Prussian blue staining to label the distribution of Fe inside tumor cells. 4T1 cells were cultured on U35 mm petri dishes for 24 h and then treated with DOX@Gd-MFe 3 O 4 NPs (100 lg/mL and 200 lg/mL) for 4 h with and without a magnet. Then, the cells were washed carefully with PBS buffer and fixed with 4% PFA solution. Subsequently, the cells were subjected to Prussian blue staining for 15 min and nuclear fast red staining for 5 min, respectively. Finally, the samples were observed by an optical microscopy to image the distribution of Fe inside cells.

In vitro combined therapy of DOX@Gd-Fe 3 O 4 NPs
The in vitro combined therapeutic performance was firstly evaluated on 4T1 cells by live/dead cell staining. 4T1 cells were seed on U35 mm petri dishes overnight and then treated with DOX, Gd-MFe 3 O 4 NPs and DOX@Gd-MFe 3 O 4 NPs (0.2 mg/mL) in the presence and absence of magnetic field. Afterwards, the cells were rinsed cleanly and exposed to NIR laser (1.5 W/cm 2 ) for 5 min. Subsequently, cells were stained with calcein AM and PI to label the live and dead cells after various treatments. Finally, the samples were imaged by a fluorescent microscope.
Afterwards, the synergistic anticancer effect of DOX@Gd-MFe 3 O 4 NPs was further quantitively assessed by the MTT assay. 4T1 cells were seed in 96-well plate overnight. Then, DOX, Gd-MFe 3 O 4 NPs and DOX@Gd-MFe 3 O 4 NPs of various concentrations were separately added to each well and coincubated with cells with and without a magnet for 24 h. Thereafter, some groups of the cells were subjected to NIR irradiation (1.5 W/cm 2 ) for 5 min and further incubated for another 24 h. Finally, the MTT assay was conducted to evaluate the surviving cells in each well. The cell viability could be calculated as the percentage of surviving cells as means of triplicate tests.

T1/T 2 relaxivity evaluation
The T 1 /T 2 -weighted relaxation studies of DOX@Gd-MFe 3 O 4 NPs were conducted on a 3.0 T GE Discovery 750 W MR scanner. The DOX@Gd-MFe 3 O 4 NPs were dispersed in 0.5% agarose gel at various Gd/Fe concentrations. The T 1 and T 2 phantom images and relaxation time were obtained separately using the spin-echo sequence. The parameters for T 1weighted MRI could be set as: TR¼ 425 ms, TE ¼ Min Full, matrix size ¼ 384 Â 224, field of view¼ 18 cm Â 18 cm, slice thickness¼ 3.0 mm, and spacing ¼ 1.5 mm. The T 2 -weighted MR phantom images and relaxation times were acquired according to following parameters: repetition time (TR) ¼ 3000 ms, echo time (TE) ¼ 10 ms.

In vivo MR tumor imaging
Balb/c mice were supplied by the Animal Center of Xuzhou Medical University and the animal studies were approved by the Animal Care Committee of Xuzhou Medical University. For the tumor model development, 4T1 cells (2 Â 10 6 cells in 100 mL) were subcutaneously injected on the back of each mouse. Thereafter, after the tumor grew to about 100 mm 3 , the mice were intravenously injected with DOX@Gd-MFe 3 O 4 NPs in with and without a magnet at the tumor site during the experiment. Afterwards, T 1 /T 2 -weighted MR imaging studies were carried out at the determined time points (0, 1 h, 2 h, 3 h, 4 h, 8 h, 12 h, and 24 h) on the previous MR scanner with a special coil designed for small animal imaging. NPs þ NIR þ magnet (10 mg/kg). The mice were injected with the above formulations via tail vein and then subjected to NIR laser irradiation (2 W/cm 2 , 5 min) after 24 h injection for the NIR-treated groups. The temperature change of the tumor site was recorded by the previous IR thermal imaging system. During the treatment period, tumor volume and body weight were monitored every other day for 3 weeks.

In vivo combined antitumor activity
The tumor volume could be calculated according to the equation as: V ¼ (length of tumor) Â (width of tumor) 2 /2.

In vivo biosafety study
The long-term biosafety study was conducted on healthy Balb/c mice for a 21-d period. Balb/c mice were (7 weeks) were randomly separated into two groups and injected with saline and Gd-MFe 3 O 4 NPs (10 mg/kg). Blood samples were collected at 0, 1, 7, and 21-d after injection from the mice for blood routine and biochemistry analysis. Afterwards, major organs were harvested, fixed with 10% formalin, and subjected to histological analysis. The tissue histological samples were imaged using a microscope.

Statistical analysis
Student's t-tests were used to assess the statistical significance between two groups. The p value less than .05 was considered statistically significant. All results in this work were presented as mean value ± standard deviation.

Characterization of DOX@Gd-MFe 3 O 4 NPs
In this study, DOX@Gd-MFe 3 O 4 NPs were constructed by Gddoped mesoporous Fe 3 O 4 nanoparticles following with the doxorubicin (DOX) loading in the mesopores of the NPs for magnetic targeted T 1 -T 2 dual MR imaging guided combined cancer therapy. First, mesoporous Fe 3 O 4 NPs were synthesized by the hydrothermal method. The size and the morphology of MFe 3 O 4 NPs were explored by TEM as shown in Figure 1(A). The MFe 3 O 4 NPs exhibited uniform spherical shape with an average diameter of about 49 nm and clear mesoporous nanostructure (Figure 1(A,C)). BET analysis was also deployed to investigate the mesoporous structures of the MFe 3 O 4 NPs. From the nitrogen adsorption-desorption in Figure 1(D), the surface area of the was evaluated to be 12.3 m 2 g À1 , which was enough for the drug encapsulation. The pore size analysis indicated the size distribution in the range of 3.52-5.87 nm with a peak pore size of 4. 7 nm. The crystal structure of the MFe 3 O 4 NPs was explored by the X-ray diffraction (XRD) analysis. As shown in Figure 1(E), the synthesized MFe 3 O 4 NPs revealed the X-ray diffraction peaks   (Figure 1(F)). The X-ray photoelectron spectroscopy (XPS) analysis also revealed characteristic peaks of Fe (2p), O(1s), and Gd (4d) elements (Figures 1(G) and Figure S1). Figure 1(H) reveals that DOX@Gd-MFe 3 O 4 NPs could be readily suspended in DI water to form a stable black dispersion in an aqueous solution. And the DOX@Gd-MFe 3 O 4 NPs could be drawn to sidewall under an external magnet. The magnetic properties of MFe 3 O 4 NPs were evaluated by a vibrating sample magnetometer (VSM) and analyzed as 112 emu/g, indicating their high magnetization of MFe 3 O 4 NPs. In addition, the magnetization of DOX@Gd-MFe 3 O 4 NPs slightly decreased to 98 emu/g attributed to the existence of non-magnetic components in the DOX@Gd-MFe 3 O 4 NPs. In addition, we also investigated the UV-vis-NIR absorbance of these nanomaterials. As shown in Figure 1(I), the MFe 3 O 4 NPs exhibited broad and continuous absorbed spectra in the NIR range. After DOX loading, a clear absorbance peak at 480 nm corresponding to DOX appeared in the absorbance of DOX@Gd-MFe 3 O 4 NPs. Moreover, the loading efficiency of DOX in MFe 3 O 4 NPs was explored and calculated to 10.5%. Thus, these results confirmed the successful synthesis of DOX@Gd-MFe 3 O 4 NPs from multiple perspectives.

Photothermal effect of DOX@Gd-MFe 3 O 4 NPs
The photothermal conversion is of great significance for the light-absorbing agent to be applied in PTT. The photothermal property of the DOX@Gd-MFe 3 O 4 NPs with different concentrations was performed by exposing to NIR laser at various power intensities. As shown in Figure 2(A-C), upon the NIR irradiation, the nanoparticles solutions were rapidly heated to a plateau temperature and exhibited concentration-dependent temperature elevation in 5 min. The temperature of the particle solutions increased to from 35.7 C to 69.9 C as the concentrations ranged from 0.05 to 1 mg/mL. In contrast, the temperature of PBS only increased by 4.3 C in the same condition. It demonstrated that the DOX@Gd-MFe 3 O 4 NPs possessed good photothermal property to convert NIR laser into thermal energy. Subsequently, we further investigated the photothermal performance of DOX@Gd-MFe 3 O 4 NPs under various NIR laser power intensities. As expected, as revealed in Figure 2(D), a greater NIR power density induced higher temperature elevation. The NPs solution could be easily heated to 45 C at a low power density of 1 W/cm 2 , which could lead to irreversible damage to tumor cells. Moreover, the photothermal conversion efficiency of DOX@Gd-MFe 3 O 4 NPs was measured and calculated to be 26.8% (Supplementary data). Furthermore, the photothermal conversion stability of DOX@Gd-MFe 3 O 4 NPs was also investigated by repeatedly irradiating the DOX@Gd-MFe 3 O 4 NPs solutions. As shown in Figure 2(E), the DOX@Gd-MFe 3 O 4 NPs solutions exhibited identical performance in the four consecutive heating and cooling cycles, demonstrating the good photothermal stability of the DOX@Gd-MFe 3 O 4 NPs. These results indicated the superior light-absorbing ability of DOX@Gd-MFe 3 O 4 NPs for PTT application.

In vitro DOX release from DOX@Gd-MFe 3 O 4 NPs
The drug release characteristics of DOX@Gd-MFe 3 O 4 NPs were studied based on the fluorescence of DOX in the released medium by a dialysis method. In addition, we also investigated the influence of NIR irradiation on the DOX release. As shown in Figure 3, as a result, the DOX@Gd-MFe 3 O 4 NPs exhibited relatively low release rate of 31%, attributed the holding effect of mesopore of MFe 3 O 4 NPs. The small amount of DOX release might be attributed the leakage of DOX from MFe 3 O 4 NPs. However, under NIR irradiation, the DOX release was significantly activated, where burst release behavior was observed shortly after NIR irradiation. Furthermore, the overall DOX release rate was accelerated under NIR irradiation. And at the end, almost 79% of DOX was liberated into the medium. It may be attributed to the local high temperature induced by the photothermal effect. These results demonstrated that NIR irradiation could be used to trigger the DOX release from DOX@Gd-MFe 3 O 4 NPs to enable on-demand drug release specifically at tumor site.

T1 longitudinal and T 2 transverse relaxivity
In this work, Gd was doped on the MFe 3 O 4 NPs to form the Gd-MFe 3 O 4 NPs for the simultaneous dual T 1 and T 2weighted MR imaging. The longitudinal and transverse relaxivities of Gd-MFe 3 O 4 NPs were investigated by the in vitro MR phantom tests at various concentrations dispersed in 0.5% agarose gel. As shown in Figure 4(A), the MR phantom images clearly revealed increasing brightening and darkening effect dependent on the increasing Gd and Fe concentrations, respectively. It confirmed the simultaneous T 1 -and T 2weighted dual contrast potential of Gd-MFe 3 O 4 NPs in MR imaging. Furthermore, the T 1 longitudinal and T 2 transverse relaxivity was calculated to quantitatively evaluate the dual T 1 and T 2 -weighted MR contrast ability of Gd-MFe 3 O 4 NPs. Figure 4(B,C) shows the linear plots of longitudinal and transverse relaxation rates (r 1 and r 2 ) vs the concentrations of Gd and Fe in the samples, respectively. The Gd-MFe 3 O 4 NPs revealed an enhanced r 1 relaxivity of 9.64 mM À1 s À1 compared to the commercial product Gd-DTPA of 4.45 mM À1 s À1 . On the other side, the r 2 rate was found to be 177.71 mM À1 s À1 , which was much higher than many previous reports (Chen et al., 2011;Zhou et al., 2014). The high r 1 and r 2 rates would enable the Gd-MFe 3 O 4 NPs in efficient dual T 1 and T 2 -weighted MR imaging.

In vitro biocompatibility of Gd-MFe 3 O 4 NPs
The biological safety is critical for nanomaterials to be used in biomedical fields, such as cancer diagnosis and therapy.
The biocompatibility of Gd-MFe 3 O 4 NPs was first evaluated in vitro on NIH3T3 and 4T1 cells by MTT assay, as shown in Figure S2. Both NIH3T3 and 4T1 cells exhibited good growth condition after the 24 h treatment of Gd-MFe 3 O 4 NPs. And NIH3T3 and 4T1 cells remained high cell viabilities of 92% and 86%, respectively, at the highest concentration of 0.5 mg/mL. This result preliminary confirmed the biocompatibility of Gd-MFe 3 O 4 NPs in vitro.

In vitro cellular uptake
The cellular uptake efficacy is of great importance for the targeted intracellular delivery of nanomedicine for tumor therapy. In this study, cellular uptake was explored using CLSM and Prussian blue staining on 4T1 cells. In addition, we also investigated the influence of NIR irradiation and magnetic field to the internalization of DOX@Gd-MFe 3 O 4 NPs. For CLSM observation, 4T1 cells were cultured on sterile coverslips and incubated with DOX@Gd-MFe 3 O 4 NPs for 2 h with and without a magnet. A group of the samples was exposed to NIR laser (1.5 W/cm 2 , 5 min) for 5 min after 1 h incubation of the NPs. As shown in Figure 5   were engulfed inside 4T1 tumor cells with relatively low red fluorescence surrounding the blue cell nuclei. It may be attributed to low targeting ability and intracellular DOX release from DOX@Gd-MFe 3 O 4 NPs. Subsequently, in the presence of magnetic field, the DOX signal was highly enhanced compared to the previous group, demonstrating a magnetic-field guided intracellular delivery of magnetic DOX@Gd-MFe 3 O 4 NPs to tumor cells. Furthermore, we also deployed NIR laser to boost the tumor cellular uptake and trigger intracellular DOX release. After NIR irradiation, the DOX signal was much higher inside cells than the other groups. The result may be partly ascribed to the improved membrane permeability of tumor cells after NIR irradiation. The other reason is NIR laser could efficiently trigger the release of DOX from DOX@Gd-MFe 3 O 4 NPs.
The behavior of internalization of DOX@Gd-MFe 3 O 4 NPs was also measured by Prussian blue staining to directly determine the distribution of DOX@Gd-MFe 3 O 4 NPs inside cells ( Figure 5(B)). As a result, compared to the control groups, many blue spots were observed inside and attached to the tumor cells after co-incubation. In addition, much higher internalization of DOX@Gd-MFe 3 O 4 NPs into cells was clearly confirmed under the guiding of magnetic field, which was consistent to the CLSM result. These results indicated the magnetic targeted tumor intracellular delivery of DOX@Gd-MFe 3 O 4 NPs and NIR triggered release inside cells.

In vitro combined therapy of PTT and chemotherapy
Inspired by the superior photothermal performance of DOX@Gd-MFe 3 O 4 NPs, we moved to investigate the anticancer performance of PTT combined with chemotherapy on 4T1 cells. Live/dead staining was first performed to evaluate the therapeutic outcome of combined therapy with Calcein AM and PI, and then observed by fluorescence microscope. As shown in Figure 6(A), the cells in control groups were not significantly influenced and remained good condition, suggesting Gd-MFe 3 O 4 NPs possessed limited damage to tumor  cells. For the DOX@Gd-MFe 3 O 4 NPs group, a small part of cells was found to be dead owing to the released DOX inside tumor cells. In contrast, for the Gd-MFe 3 O 4 NPs group upon NIR irradiation, cells were dramatically killed by the photothermal effect, indicating the excellent killing effect of the PTT only. Furthermore, when combined with DOX, the DOX@Gd-MFe 3 O 4 NPs þ NIR group exhibited even much more cells stained to be red with a few green fluorescence, suggesting the advantage of the combined therapy. Particularly, when we applied magnetic field to facilitate targeting of DOX@Gd-MFe 3 O 4 NPs, all 4T1 cells were almost killed, indicating the therapeutic efficacy was remarkably enhanced by the improved cellular uptake mediated by external magnetic field. The tumor therapeutic performance was further measured quantitatively by the MTT assay. As exhibited in Figure  6(B,C), NIR only and Gd-MFe 3 O 4 NPs treated groups remained high cell viabilities above 90% compared to the control group. Free DOX induced similar killing effect of 54% to tumor cells with and without NIR laser. However, the DOX@Gd-MFe 3 O 4 NPs groups with and without a magnet revealed limited anticancer effect due to the low release of DOX from DOX@Gd-MFe 3 O 4 NPs inside tumor cells. However, under NIR irradiation, the cell viabilities of DOX@Gd-MFe 3 O 4 NPs ± magnet were reduced to 13% and 24%, respectively, attributed to the synergistic effect of PTT and DOX. Besides, the cell viability was highly dependent on the concentration of different treatments of Gd-MFe 3 O 4 NPs, DOX@Gd-MFe 3 O 4 NPs, and DOX@Gd-MFe 3 O 4 NPs þ magnet under NIR irradiation. Especially, when the concentration increased to 0.25 mg/mL, cell viability of the DOX@Gd-MFe 3 O 4 NPs þ magnet group significantly decreased to 15.2%. These results suggested a good therapeutic outcome of the combined PTT and DOX therapy for cancer treatment.

In vivo magnetic targeted MR imaging
Considering the high r 1 and r 2 relaxivity of the prepared DOX@Gd-MFe 3 O 4 NPs, we further performed T 1 and T 2 weighted MR imaging on 4T1 bearing mice. The DOX@Gd-MFe 3 O 4 NPs were administrated intravenously and then the mice were subjected to MR imaging at various time points. As shown in Figure 7(A), as expected, T 1 signal of the tumor region became bright after 1 h injection and increased to the peak at 4 h. Thereafter, the T 1 signal gradually decayed and almost disappeared after 24 h. To investigate the effect of magnetic field to the NPs accumulation, a magnet was also attached at the tumor site of each mouse of a group. As revealed in Figure 7(A), the MR signal was generally enhanced in all time points after magnet attachment. Especially, the MR intensity was extremely strengthened at 3 h and maintained a very high signal until 8 h compared to 1 h high intensity window period (3-4 h after injection) for the group without magnet attachment. It suggested that external magnetic field could highly improve the accumulation of the DOX@Gd-MFe 3 O 4 NPs at tumor site. As a result, the MR signal at tumor region was strengthened due to the existence of large amount of DOX@Gd-MFe 3 O 4 NPs. Moreover, T 1 MR signal intensity was also measured at the tumor region at the determined intervals. As shown in Figure 7(C), MR signal of the two groups gradually increased to the peak at 4 h and then decayed over time, which was consistent to the MR images. Particularly, for the peak signal at 4 h, the magnet attached group was 1.6-fold compared to the group without magnet attachment. Furthermore, T 2weighted MR imaging was also performed at 4 h post-injection. For the control group, the tumor region exhibited no difference compared to muscle tissues. However, obvious tumor darkening effect was found at the tumor region upon the injection of DOX@Gd-MFe 3 O 4 NPs (Figure 7(B)). Furthermore, in the presence of magnetic field, the tumor region became much darker compared to the group without external magnetic field. Consistently, T 2 MR signals in tumors decreased to 0.46-and 0.26-folds for mice without and with a magnet attached, respectively compared to the control group ( Figure S3). These results demonstrated the feasibility of DOX@Gd-MFe 3 O 4 NPs for magnetic targeted dual T 1 /T 2 MR imaging in specific cancer diagnosis.

In vivo anticancer outcome of combined PTT and chemotherapy
Encouraged by the superior tumor therapeutic effect in vitro, we next evaluated the dual MR imaging-guided combined PTT and chemotherapy in vivo. The temperature elevation in tumor area upon NIR irradiation was monitored by an IR thermal camera, as shown in Figures 8(A) and Figure S4. For mice injected with saline, the tumor region merely increased by 4.1 C under the NIR irradiation (1.5 W/cm 2 ) for 5 min, suggesting the negligible effect of the NIR laser only. In contrast, for the mice receiving DOX@Gd-MFe 3 O 4 NPs and MFe 3 O 4 NPs, the laser induced rapid and significant temperature elevation at 17.5 and 18.2 C, respectively, demonstrating the excellent ability of the prepared NPs to convert laser into thermal energy in vivo. In addition, as expected, when we attached a magnet at the tumor site, the mice revealed highest temperature increase of 22.7 C at tumor area attributed to the magnetic field mediated tumor targeting. Subsequently, the tumor volumes of each mouse after treatments were monitored for 14 days by a caliper. Afterwards, the result was presented as relative tumor volume compared to the initial volume, as shown in Figure 8(B). For the saline and NIR laser only treated mice, the tumor rapidly grew to 9.1-and 7.9-folds at the end, respectively, where the low temperature elevation at tumor site had limited tumor inhibition. On the other side, mice receiving free DOX, DOX@Gd-MFe 3 O 4 NPs exhibited moderate tumor inhibition by 5.7-and 6.9-folds of initial volume, indicating that mono-chemotherapy could not efficiently suppress the growth of tumor in a long term. In contrast, the tumor growth was highly inhibited to 4.5-fold for Gd-MFe 3 O 4 NPstreated mice receiving NIR laser irradiation, demonstrating the effectiveness of mono-PTT in tumor inhibition. Furthermore, DOX@Gd-MFe 3 O 4 NPs exhibited much higher tumor suppression to 0.4-fold under NIR laser irradiation. The therapeutic outcome was further enhanced when we applied external magnetic field to guide the accumulation of DOX@Gd-MFe 3 O 4 NPs at the tumor site. In addition, mice and tumor photographs further verified the real therapeutic outcome of various treatments (Figure 8(C,D)). Moreover, no mice were found dead or body weight loss during the test period, revealing low effect of the formulations to mice ( Figure S5). These results demonstrated the advantage of magnetic field guided combined PTT and chemotherapy in cancer treatment.

In vivo toxicity analysis
The in vivo long-term biosafety was investigated on healthy Balb/c mice in a 21-day period. Gd-MFe 3 O 4 NPs in saline were intravenously administrated in mice. Then, blood samples were collected from each mouse at 0, 1, 7, and 21 day post-injection for blood routine and biochemistry analysis. Several indicators were evaluated for the blood routine tests, including white blood cells (WBC), red blood cells (RBC), hemoglobin (HGB), platelet (PTL), mean corpuscular hemoglobin (MCH), mean corpuscular volume (MCV), mean corpuscular hemoglobin concentration (MCHC), and hematocrit (HCT). As show in Figure S6, all indicators in days 1, 7, and 21 did not reveal obvious variation compared with the control group in day 0. The serum biochemistry analysis was also conducted in the same condition. As revealed in Figure  9(A), the values of aspartate aminotransferase (AST), alanine aminotransferase (ALT), total protein (TP), indicatorsalbumin (ALB), blood urea nitrogen (BUN), and creatinine slightly fluctuated during the test period, where the values were still within the normal ranges for each indicator.
Furthermore, we also evaluated the biosafety of Gd-MFe 3 O 4 NPs by the histological tissue analysis. Major organs were harvested after we withdrew blood samples, sliced, and stained with hematoxylin-eosin. As shown in Figure 9(B), the tissues in days 1, 7, and 21 revealed no obvious difference with the control group in day 0. Not only that, no remarkable inflammatory lesion and organ damage were observed in the experimental groups. These results confirmed the good biosafety of MFe 3 O 4 NPs in vivo in a long term, which would be compatible for biomedical applications.

Conclusions
In summary, we successfully synthesized a magnetic targeted nanoplatform DOX@Gd-MFe 3 O 4 NPs for T 1 /T 2 dual modal MR imaging-guided synergistic cancer therapy. The prepared DOX@Gd-MFe 3 O 4 NPs exhibited good colloidal dispersity, superior magnetic properties, superior NIR photothermal conversion, and NIR-triggered DOX release. High T 1 /T 2 MR contrast was confirmed on the DOX@Gd-MFe 3 O 4 NPs, which favored specific T 1 /T 2 dual-modal MR imaging on 4T1 bearing mice in vivo. More than that, as a result of the good NIR photothermal conversion, DOX@Gd-MFe 3 O 4 NPs demonstrated remarkably improved therapeutic effect of combined PTT and chemotherapy to eliminate cancer. Moreover, the blood and histological analysis validated the good Figure 9. In vivo toxicity tests of Gd-MFe 3 O 4 NPs: (A, a-f) mouse serum biochemistry analysis before (0d, control) and after injection of NPs for 1, 7, and 21 days. (B) Histological images of the heart, liver, spleen lung, and kidney of healthy mice 1, 7, and 21 days injected with Gd-MFe 3 O 4 NPs and control mice. The organs were sectioned and stained with hematoxylin and eosin (H&E) and observed under a light microscope, scale bar ¼ 100 lm. biocompatibility of the NPs. Therefore, this versatile nanoplatform would provide a promising strategy for T 1 /T 2 dualmodal MR imaging guided synergistic cancer therapy.

Disclosure statement
There are no conflicts of interest.