Folic acid-targeted iron oxide nanoparticles as contrast agents for magnetic resonance imaging of human ovarian cancer

Background Improved methods for the early and specific detection of ovarian cancer are needed. Methods In this experimental study, we used folic acid (FA)-targeted iron oxide (Fe3O4) nanoparticles (NPs) as a T2-negative contrast agent for magnetic resonance (MR) imaging to accurately detect ovarian cancer tissues in an intraperitoneal xenograft tumor model. Human serous ovarian cell line (Skov-3), with overexpressed FA receptors, was chosen as the targeted tumor cell mode. For in vivo experiments, the cells were injected intraperitoneally into nude mice to produce intraabdominal ovarian cancers. FA-targeted and non-targeted Fe3O4 NPs were prepared. Results FA-targeted Fe3O4 NPs with a mean size of 9.2 ± 1.7 nm have a negligible cytotoxicity to human serous ovarian cell line (Skov-3). Importantly, the results of cellular uptake suggested that FA-targeted Fe3O4 NPs have a targeting specificity to Skov-3 cells overexpressing FA receptors. FA-targeted Fe3O4 NPs could be specifically localized by magnetic resonance (MR) imaging to the intraperitoneal human ovarian carcinoma tissues, as documented by a statistically significant difference (p = 0.002, n = 3) in T2 signal intensities of xenograft tumor tissues when injected with FA-targeted and non-targeted Fe3O4 NPs at 4 h post-injection. Conclusion FA-targeted Fe3O4 NPs appear to be promising agents for the detection of human ovarian carcinoma by MR imaging, and possibly also for the hyperthermal treatment of the tumors.


Background
Ovarian cancer is the sixth most commonly diagnosed cancer in the world, accounting 4 % of all cancers in women [1], and it is the leading cause of death from gynecologic malignancies in the western world [2,3]. Most ovarian cancers are first diagnosed in an advanced stage because patients' symptoms may be minimal or nonspecific and no reliable biomarkers are available [4]. Tumor-debulking surgery is the first choice of management for most patients with ovarian cancer [5], but most ovarian cancers recur after surgery and are intractably drug resistant [6]. Therefore, although some advances in cytoreductive surgery and case-effective chemotherapy have been made in the last decade, the prognosis for ovarian cancer, especially for epithelial ovarian cancer still is limited.
In most tertiary medical centers, magnetic resonance (MR) imaging is generally performed for imaging assessment of complex ovarian masses [7,8] that are indeterminate on either palpation or ultrasonography because of MR's superb soft-tissue resolution and lack of radiation. The MR diagnostic criteria for ovarian malignancies are based on morphology: thick septum, vegetations, ascites, lymphadenopathy, and vividly enhancing solid component, which are features well described in numerous reports [8,9]. However, identification of the tumor tissues at an early stage with available imaging modalities still possesses a great challenge for both radiologists and clinicians.
Recent advances in nanoscience and nanotechnology have enabled the development of various contrast agents for MR imaging applications, such as Gd (III)-or Mn (II)-based T 1 MR contrast agents [10,11] and magnetic iron oxide nanoparticle (Fe 3 O 4 NPs)-based T 2 MR contrast agents [12][13][14]. The Fe 3 O 4 NPs are the most commonly used magnetic materials for various biomedical applications [15][16][17][18]. But, few reports on the application of Fe 3 O 4 NPs for the diagnosis of ovarian cancer have been published.
Folic acid (FA) receptors as single-chain glycoproteins with high specific affinity for FA are highly overexpressed on various malignant tumors, including human ovarian cancer [19]. The over-expression of FA receptors on malignant tumor tissues can be exploited as a specific targeting ligand since most healthy tissues have little FA receptors expression [20]. This targeting strategy has the potential for diagnostic and therapeutic application in a wide variety of cancers [21,22].
In this research, we used FA-targeted Fe 3 O 4 NPs as T 2 -negative contrast agents for in vivo MR imaging of ovarian cancer in an intraperitoneal xenograft tumor model. To the best of our knowledge, this is the first reported application of FA-targeted Fe 3 O 4 NPs in MR imaging diagnosis of ovarian cancer.

Synthesis and characterization techniques
FA-targeted Fe 3 O 4 NPs were synthesized and characterized according to our previous work [23]. Non-targeted Fe 3 O 4 NPs were synthesized by the same methods, except for the use of mPEG-COOH in the PEGylation step instead of FA-PEG-COOH.
Branched polyethyleneimine (PEI, Mw = 25,000)-coated Fe 3 O 4 NPs (Fe 3 O 4 @PEI NPs) were synthesized via a reduction route. FeCl 3 · 6H 2 O (1.3 g) was dissolved in 20 mL water, and placed into a 250 mL three-necked flask. Under vigorous stirring, the solution was bubbled with nitrogen atmosphere for 15 min, then 10 mL freshly prepared sodium sulfite solution (0.2 g) was added slowly into the flask. 30 min later, 5 mL PEI (0.5 g) and 2 mL ammonia (25 %) was added into the flask successively. The reaction mixture was vigorously stirred for 30 min at 60~70 o C, and then at room temperature for another 1.5 h. The product (Fe 3 O 4 @PEI NPs) was magnetically collected and washed 3 times with water. Finally, the sample was centrifuged (8000 rpm, 10 min) to remove the aggregation and larger particles.
An aqueous solution of Fe 3 O 4 @PEI NPs (110 mg, 35 mL) was precipitated by virtue of an external magnet and re-dispersed in 20 mL DMSO. Another solution of 38.5 mg activated FA-PEG-COOH or mPEG-COOH in 2 mL DMSO was added dropwise into the above DMSO solution of Fe 3 O 4 @PEI NPs and kept shaking for 3 d.
The formed products were collected by magnetic separation and washed with DMSO for 3 times to remove excess reactants. Finally, the amino groups on the surface of the particles were acetylated by reaction with acetic anhydride. Briefly, triethylamine (493 μL) was added into the aqueous solution of raw product of Fe 3 O 4 @PEI-PEG-FA NPs or Fe 3 O 4 @PEI-mPEG NPs under vigorous shaking using a shaker at room temperature. After 30 min, acetic anhydride (402 μL) was dropwise added into the above mixture solution and the reaction was continued for 1 d. After several times magnetic separation/washing/dispersion steps to remove excess reactants and by-products, the final products (FA-targeted Fe 3 O 4 NPs and non-targeted Fe 3 O 4 NPs) were obtained, re-dispersed in water and stored under 4 o C for further use.
A JEOL 2010 F transmission electron microscopy (TEM, JEOL, Tokyo, Japan) was used to characterize the morphology of the FA-targeted Fe 3 O 4 NPs and nontargeted Fe 3 O 4 NPs at an operating voltage of 200 kV. A dilute particle suspension of the sample in water (10 μL) was deposited onto a carbon-coated copper grid and dried in air before measurements. The effect of MR imaging for FA-targeted and non-targeted Fe 3 O 4 NPs was evaluated with a 1.5 Tesla MR imaging machine (Siemens Avanto, Erlangen, Germany). Samples were diluted with water to have different Fe concentrations in the range of 0.005-0.08 mM before measurements. The T 2 -weighted imaging parameters with turbo spin echo sequence were set as follows: point resolution = 156 mm × 156 mm, section thickness = 1.5 mm, TR = 4000 ms, TE = 85 ms, bandwidth (Hz) = 260, number of excitation = 1, and voxel size = 1.1 × 1.1 × 4.0 mm.

Cytotoxicity of FA-targeted Fe 3 O 4 NPs and non-targeted Fe 3 O 4 NPs
The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) viability assay was carried out to evaluate the cytotoxicity of the FA-targeted Fe 3 O 4 NPs and non-targeted Fe 3 O 4 NPs. Briefly, 1 × 10 4 Skov-3 cells were seeded into each well of 96-well cell culture plates with 200 μL regular RPMI-1640 medium and cultured at 37°C and 5 % CO 2 overnight to bring the cells to confluence. Next, the medium in each well was discarded carefully and 200 μL of fresh medium containing phosphate-buffered saline (PBS), FA-targeted Fe 3 O 4 NPs or non-targeted Fe 3 O 4 NPs at the Fe concentration of 0.5 to 1.0 mM was added. After 24 h incubation at 37°C and 5 % CO 2 , 20 μL MTT solution (5 mg/mL in PBS buffer) were added to each well to reveal the viable cells. After further incubation for 4 h at 37°C and 5 % CO 2 , the medium was carefully removed, and DMSO (200 μL) was added to dissolve the formazan grains. The absorbance value of each well was measured with a microplate reader at 450 nm wavelength.

Cellular uptake of FA-targeted Fe 3 O 4 NPs and non-targeted Fe 3 O 4 NPs
To qualitatively confirm the cellular uptake of Fe 3 O 4 NPs by Skov-3 cells, the cells was stained with Prussian blue. In brief, 5 × 10 5 cells were seeded into each well of 24-well cell culture plates. After overnight incubation at 37°C and 5 % CO 2 to bring the cells to 80 % confluence, the medium was replaced with fresh medium containing PBS buffer (control), FA-targeted The Leeman Prodigy inductively coupled plasma-optical emission spectroscopy (ICP-OES, Hudson, NH, USA) also was used to quantify the cellular uptake of the Fe 3 O 4 NPs by Skov-3 cells. The Skov-3 cells were seeded into 12-well plates with a density of 1 × 10 6 cells/well. After overnight incubation to bring the cells to confluence, the medium was discarded carefully, and 1 mL fresh medium containing PBS buffer (control), FA-targeted Fe 3 O 4 NPs or nontargeted Fe 3 O 4 NPs at Fe concentrations of 0.2 and 0.4 mM was added. The cells were further incubated at 37°C and 5 % CO 2 for 4 h. The medium was then removed. The cells were washed with PBS buffer four times, trypsinsized, collected, and suspended in 1 mL PBS buffer. The cell numbers in each sample were estimated with a hemocytometer. For the cellular uptake assay, the cells were centrifuged (1000 rpm, 5 min), collected, and lysed with an aqua regia solution (0.5 mL) for 12 h. The Fe content was determined by ICP-OES after the samples were diluted 2 times with PBS.

Statistical analysis
Quantitative data were expressed as mean ± standard deviation (SD). Means were compared by use of unpaired two-sided Student's t-test. The data are indicated with (*) for p < 0.05, (**) for p < 0.01 and (***) for p < 0.001.

Synthesis and characterization techniques
The morphology of the FA-targeted Fe 3 O 4 NPs and nontargeted Fe 3 O 4 NPs was characterized by TEM (Fig. 1). It can be seen that the NPs with a spherical or quasispherical shape have a quite uniform size distribution and a polymer shell on their outer surface. The mean size was measured to be 8.7 ± 1.9 nm for non-targeted Fe 3 O 4 NPs (Fig. 1a) and 9.2 ± 1.7 nm for FA-targeted Fe 3 O 4 NPs (Fig. 1b), respectively. The T 2 -weighted MR effect of the NPs was evaluated by use of a 1.5 T MR system. Fe 3 O 4 NPs decreased the MR signal intensity in relation to increasing Fe concentration for both FA-targeted Fe 3 O 4 NPs and non-targeted Fe 3 O 4 NPs (Fig. 2). The T 2 signal intensities of FA-targeted Fe 3 O 4 NPs at the given Fe concentrations were 1851 ± 14, 1808 ± 18, 1648 ± 30, 1628 ± 71, and 1395 ± 73. The T 2 signal values in nontargeted Fe 3 O 4 NPs at each given Fe concentration were 2094 ± 28, 1838 ± 14, 1742 ± 10, 1667 ± 2 and 1487 ± 26, respectively (Fig. 3). Based on the measured T 2 relaxation time, the r 2 relaxivity of FA-targeted Fe 3 O 4 NPs and non-targeted Fe 3 O 4 NPs was calculated to be 475.92 and 545.70 mM −1 s −1 , respectively according to our previous works [23].

Cytotoxicity assay of FA-targeted Fe 3 O 4 NPs and non-targeted Fe 3 O 4 NPs
It is important to assess the potential cytotoxicity of

In vivo targeted MR imaging
After intravenous injection of the FA-targeted Fe 3 O 4 NPs or non-targeted Fe 3 O 4 NPs into the mice bearing intraperitoneal ovarian tumors, MR scanning was performed. The tumor MR signal for the mice injected with both particles gradually decreased with time after injection (Fig. 7). In the MR images, we can see the contrast enhancement was highest at 2 h post injection. After that, the tumor signal recovered because of further metabolism. Quantitative analysis of the T 2 signal intensity of solid tumors at various time points revealed that the lowest signal intensity occurred at 2 h after injection with both the FAtargeted Fe 3 O 4 NPs and non-target Fe 3 O 4 NPs (Fig. 8). The T 2 -weighted signal intensity of the lesions at 0.5, 1, 2, and 4 h post injection was 1666 ± 152, 1534 ± 92, 749 ± 56 and 1402 ± 102 for the FA-targeted Fe 3 O 4 NPs group and 1414 ± 42, 1328 ± 162, 1181 ± 93 and 1615 ± 84 for the non-targeted Fe 3 O 4 NPs group, respectively (Fig. 8). It should be noted that the T 2 signals intensity of the mice treated with FA-targeted Fe 3 O 4 NPs was significantly lower than that of the mice treated with non-targeted Fe 3 O 4 NPs at 2 h post injection (P = 0.002, n = 3). This results suggested that the prepared FA-targeted Fe 3 O 4 NPs have a great potential to be used as contrast agents for targeted MR imaging to diagnosis the ovarian tumors.

Discussion
In this study, we report our preliminary experience in imaging human ovarian cancer in the xenograft tumor model by using FA-targeted Fe 3 O 4 NPs as contrast agents. Owing to the good contrast enhancement and low cytotoxicity, the FA-targeted Fe 3 O 4 NPs can detect the ovarian cancer tissues planted in the abdominal cavity of nude mice at in vivo levels. Our results indicated that FA-targeted Fe 3 O 4 NPs hold promise for being effective magnetic molecular probes for detecting tumor tissues in gynecologic cancer.
Ovarian cancer is the most malignant gynecological tumor and therefore deserves extensive basic and clinical research in the quest for early diagnostic tests and   [24][25][26][27]. Fe 3 O 4 NPs are low-toxic and eventually biodegrade to form blood hemoglobin [14], and they have been used for liver imaging since the 1900s [28]. With recent advances in nanotechnology and nanoscience [29][30][31][32], various polymers have been coated onto the surface of Fe 3 O 4 NPs to improve their stability and decrease their uptake by the reticuloendothelial system [16,33,34]. Numerous studies on application of NPs in biomedical imaging have been reported in recent decades [35][36][37][38][39][40], but few have examined application of the particles in ovarian cancer.
In our previous work, we demonstrated that FA-targeted Fe 3 O 4 NPs have good water-dispersibility, colloidal stability and fairly high relaxivity [23]. In addition, the particles have excellent hemocompatibility and cytocompatibility in the studied range of concentrations. We found that FAtargeted Fe 3 O 4 NPs had excellent binding specificity to a human cervical cancer cell line (HeLa cells) overexpressing FA receptors via an active FA targeting pathway. In the present study, we found excellent lesion targeting ability of the FA-targeted Fe 3 O 4 NPs to ovarian cancer in the T 2 -weighted MR imaging, which may be attributed to the following aspects: First, although the mean size of the FA-targeted Fe 3 O 4 NPs was small (9.2 ± 1.7 nm), the particles had a very high r 2 relaxivity coefficients (475.92 mM −1 s −1 ), which is much higher than those of other reported Fe 3 O 4 NPs [33,40]. This feature made the particles more sensitive to magnetic susceptibility effects. Second, the presence of FA on the surface of the Fe 3 O 4 NPs increased their ability to target tumor tissues. Third, the passive enhanced permeability and retention effect into solid tumors may also facilitate the specific MR imaging of tumors [36].
Human ovarian cancers are located deep in the pelvic space [41]. An ideal humanized xenograft mouse model of ovarian cancer would simulate the true microenvironment for tumor angiogenesis [24,[42][43][44]. Thus, in the present study, we implanted the tumor cells in the abdomen rather than in subcutaneous sites, believing that the intraperitoneal location would reflect the hemodynamic condition of ovarian cancer in humans-at least more accurately than would a subcutaneous site, as has been often used [10][11][12]25]. Our results corroborated this point: both targeted and non-targeted particles were evident by T 2 -weithted MR imaging at 2 h after injection in abdominal tumors compared with 1 h in subcutaneous tumors [23], perhaps because more time was needed for Fe 3 O 4 NPs to reach the deep abdominal tumors in sufficient concentration to be evident on T 2 -enhanced imaging. We must confess that the T 2 signal intensity in MR images also achieve the lowest point at 2 h post injection of non-targeted Fe 3 O 4 NPs, which may be due to the enhanced permeability and retention (EPR) effect (passive uptake) as well documented in solid tumors [14,36,37]. However, both in vitro and in vivo imaging results (as shown in Figs. 5 and. 7) proved FA-targeted   ligands can enable the tumor uptake through more active pathway, thus making the tumors look like more dark compared with non-FA targeted group.
Further, we also found that, after injection of Fe 3 O 4 NPs in the nude mice, the tumor T 2 signal intensity had reverted to pre-injection intensity after 4 h, a little earlier than that we had found previously in subcutaneous tumors [23]. We also acknowledge that the injection of suspensions of tumor cells into the mice is different from the formation of tumors in the natural environment.
Other methods for imaging detection of ovarian cancers have been described. Hensley, et al [45] described a dual MR-fluorescence molecular tomography approach, with commercially available fluorescent molecular imaging probes for the detection and quantification of tumorassociated metabolites in ovarian carcinomas in a transgenic mouse model of epithelial ovarian cancer. The authors concluded that the combination of in vivo molecular and MR imaging can effectively detect orthotopic ovarian tumors and their response to therapy [25]. In another study, Satpathy, et al [24] reported that in an orthotopic human ovarian tumor xenograft model, HER-2-targeted magnetic NPs labeled with a near infrared dye (NIR-830) were specifically delivered into primary and disseminated ovarian tumors, enabling optical and MR imaging of tumors as small as 1 mm in the peritoneal cavity. The authors designed the non-conjugated magnetic NPs with 14 ± 3.4 nm diameter and targeted-conjugated magnetic NPs with 22.9 ± 4.8 nm diameter, respectively [24]. However, they did not report the exact MR acquisition time point, which we believe is crucial for tumor imaging, especially for magnetic NPs.
Our study also has some limitations. First, by 2 weeks after injection of ovarian cancer cells into the peritoneal cavity, the tumors often had become large (average diameter about 5 mm) with isointensity signals on T 2 -weighted MR imaging making them easily detectable and distinct from surrounding tissues, which had mostly hyperintensity signals. However, tumors at an earlier stage or smaller might not be detected because of overlapping neighboring organs (such as gut, kidney, or bladder) and background tissues. Perhaps the specificity and sensitivity could be improved by the use of bimodal magnetic nanoprobes with fluorescent materials incorporated into Fe 3 O 4 NPs. Second, since FA receptors are overexpressed in most malignant tumors, the FA targeting ligand we used may not be specific for detecting ovarian cancer. Further studies should be conducted to image ovarian cancer with targeting motifs that may be more specific.

Conclusion
In summary, this study demonstrated that the prepared FA-targeted Fe 3 O 4 NPs can bound specifically in vitro to the FA receptors overexpressed human serous ovarian cells without inducing cytotoxicity. Used as T 2 -negative contrast agents in MR imaging, the particles also localized to intraperitoneal human ovarian cancer tissues in a xenograft tumor model. Importantly, the tumor can be detected more obviously after the mice were injected with FA-targeted Fe 3 O 4 NPs than non-targeted Fe 3 O 4 NPs. Thus, FA-targeted Fe 3 O 4 NPs hold promise that they may be multifunctional nanoprobes for the diagnosis and treatment of ovarian cancer.

Ethics approval
All the animal experiments were performed according to protocols approved by the Ethical Committee of Obstetrics and Gynecology Hospital ([2007]-No. 6), Fudan University.

Consent for publication
Not applicable.