Analysis of Stable Chelate-free Gadolinium Loaded Titanium Dioxide Nanoparticles for MRI-Guided Radionuclide Stimulated Cancer Treatment

Abstract

Background:

Recent studies demonstrate that titanium dioxide nanoparticles (TiO₂ NPs) are an effective source of reactive oxygen species (ROS) for photodynamic therapy and radionuclide stimulated dynamic therapy (RaST). Unfortunately tracking the in vivo distribution of TiO₂ NPs noninvasively remains elusive.

Objective:

Given the use of gadolinium (Gd) chelates as effective contrast agents for magnetic resonance imaging (MRI), this study aims to (1) develop hybrid TiO₂-Gd NPs that exhibit high relaxivity for tracking the NPs without loss of ROS generating capacity; and (2) establish a simple colorimetric assay for quantifying Gd loading and stability.

Methods:

A chelate-free, heat-induced method was used to load Gd onto TiO₂ NPs, which was coated with transferrin (Tf). A sensitive colorimetric assay and inductively coupled plasma mass spectrometry (ICP-MS) were used to determine Gd loading and stability of the TiO₂-Gd-Tf NPs. Measurement of the relaxivity was performed on a 1.4 T relaxometer and a 4.7 T small animal magnetic resonance scanner to estimate the effects of magnetic field strength. ROS was quantified by activated dichlorodihydrofluorescein diacetate fluorescence. Cell uptake of the NPs and RaST were monitored by fluorescence microscopy. Both 3 T and 4.7 T scanners were used to image the in vivo distribution of intravenously injected NPs in tumor-bearing mice.

Results:

A simple colorimetric assay accurately determined both the loading and stability of the NPs compared with the expensive and complex ICP-MS method. Coating of the TiO₂-Gd NPs with Tf stabilized the nanoconstruct and minimized aggregation. The TiO₂-Gd-Tf maintained ROS-generating capability without inducing cell death at a wide range of concentrations but induced significant cell death under RaST conditions in the presence of F-18 radiolabeled 2-fluorodeoxyglucose. The longitudinal (r1 = 10.43 mM⁻¹s⁻¹) and transverse (r2 = 13.43 mM⁻¹s⁻¹) relaxivity of TiO₂-Gd-Tf NPs were about twice and thrice, respectively, those of clinically used Gd contrast agent (Gd-DTPA; r1 = 3.77 mM⁻¹s⁻¹ and r2 = 5.51 mM⁻¹s⁻¹) at 1.4 T. While the r1 (8.13 mM⁻¹s⁻¹) reduced to about twice that of Gd-DTPA (4.89 mM⁻¹s⁻¹) at 4.7 T, the corresponding r2 (87.15 mM⁻¹s⁻¹) increased by a factor 22.6 compared to Gd-DTPA (r2 = 3.85). MRI of tumor-bearing mice injected with TiO₂-Gd-Tf NPs tracked the NPs distribution and accumulation in tumors.

Conclusion:

This work demonstrates that Arsenazo III colorimetric assay can substitute ICP-MS for determining the loading and stability of Gd-doped TiO₂ NPs. The new nanoconstruct enabled RaST effect in cells, exhibited high relaxivity, and enhanced MRI contrast in tumors in vivo, paving the way for in vivo MRI-guided RaST.

1. INTRODUCTION

Titanium dioxide nanoparticles (TiO₂ NPs) are regenerative catalysts for reactive oxygen species (ROS) production when exposed to ultraviolet (UV) light in aqueous media [1, 2]. Given that ROS induces multiple cellular cell death pathways, TiO₂ NPs are well suited for use as nanophotosensitizers for photodynamic therapy (PDT) [3]. A major limitation of this application with TiO₂ NPs is the need to excite the materials with UV light, which does not penetrate beyond superficial tissue surfaces. Recent studies are exploring an alternative approach that uses UV light-emitting Cerenkov radiation from radiopharmaceuticals or X-rays to stimulate ROS production [4, 5]. Many radiopharmaceuticals currently used in the clinic can be re-purposed for radionuclide stimulated dynamic therapy (RaST), where a photosensitizer is first administered into the body. Following selective uptake in a target tissue such as cancer, the radiopharmaceutical is then injected to generate ROS upon interaction with the photosensitizer. This approach was first demonstrated with TiO₂ NPs as a nanophotosensitizer and radiolabeled fluorine-18 deoxyglucose (FDG), achieving an excellent treatment response [6].

RaST requires good timing of when to administer the ROS-generating photosensitizer, followed by the radiopharmaceuticals. Considering that FDG is an imaging agent, its biodistribution and tumor uptake can readily be monitored over time, but TiO₂ NPs have no imaging signal to study their distribution noninvasively. As a result, animals are euthanized, and the relevant tissues are used to determine the metal concentrations using an inductively coupled plasma mass spectrometry (ICP-MS) analysis method. This process is disruptive, expensive, and time-consuming. Coating the surface of these NPs with near-infrared (NIR) fluorescent dyes is a viable approach to overcome this impediment, allowing real-time display of the NP distribution in vivo [7]. However, separation of the dye coating from the NPs in vivo gives discordance between the location of the NPs and the dye fluorescence, leading to a false representation of the NPs distribution. Another approach is to radiolabel Ti using titanium-45, which has a 3-hour half-life [8], but this approach will override the orthogonal targeting and activation strategy of RaST. Such nanomaterials will spontaneously generate cytotoxic ROS upon injection, resulting in nonspecific toxic effects and interference with positron emission signals from FDG.

Multimodality imaging is an emerging trend that harnesses the strengths of each imaging system to enhance diagnostic information. In particular, the combination of positron emission tomography (PET) and magnetic resonance imaging (MRI) offer complementary molecular and anatomical information, respectively [9]. Thus, modifying TiO₂ with Gd will enable the detection of the NPs in vivo and provide high soft tissue contrast [10, 11]. Gd chelates such as Gd-DTPA are widely used in the clinic as T1-weighted MRI contrast agents due to their high paramagnetism [10]. Previous studies have doped TiO₂ with Gd, demonstrating an enhancement of longitudinal relaxivity (r1; 12 mM⁻¹s⁻¹) compared to r1 of 3.77 mM⁻¹s⁻¹ for Gd-DTPA [12, 13]. However, there is a dearth of literature supporting the use of these agents for in vivo MRI. Part of the challenge is these doping and encapsulation methods alter ROS generation, stability, and relaxivity of the NPs, creating the need to balance stabilization of the nanomaterials with the retention of ROS generation while providing access to bulk water molecule to enhance r1 or r2 relaxivity.

Here, we report the preparation and characterization of stable Gd doped TiO₂ NPs using a chelate-free, heat-induced method. Stabilization was achieved with transferrin (Tf) coating, which allowed access to water molecules, thereby enhancing retaining ROS production and high r1 and r2 relaxivity compared to Gd-DTPA. Using a simple colorimetric assay, we quantitatively demonstrated the incorporation and stability of the NPs over time, an outcome that correlated with the complex ICP-MS analysis. A combination of low cytotoxicity, effective ROS generation under UV light, and response to RaST effect with FDG, suggests the potential use of these NPs to treat cancer. Detection of the NPs distribution in vivo points to the potential of using these materials for MRI-guided RaST.

2. MATERIALS AND METHODS

2.1. Materials

Titanium dioxide (TiO₂) anatase nanoparticles (25 nm), gadolinium chloride hexahydrate (GdCl₃·6H₂O), diethylenetriaminepentaacetic acid gadolinium (III) dihydrogen salt hydrate (Gd-DTPA), dichlorofluorescein diacetate (DCF·DA), Phosphate Buffered Saline (PBS), (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), and other reagents were purchased from Sigma Aldrich (St. Louis, MO, USA) if not stated otherwise. Human apo-Transferrin (Tf) was purchased from Athens Research & Technology (Athens, GA, USA). 2,2′-(1,8-Dihydroxy-3,6-disulfonaphthylene-2,7 bisazo) bisbenzenearsonic acid, 2,7-Bis(2-arsonophenylazo) chromotropic acid (Arsenazo III) and Transferrin from Human Serum, Alexa Fluor^™ 680 Conjugate (AlexaTf) were purchased from Thermo Fisher Scientific (Berkeley, MO, USA). Water (> 18.2 MΩ·cm at 25°C, Milli-Q, Millipore, Billerica, MA, USA) was purified by passing it through a 10 cm long column of Chelex resin (Bio-Rad Laboratories, Hercules, CA, USA) at a flow rate of < 1.0 mL/min, which removed common metal ion contaminants.

2.2. Preparation of TiO₂-Gd-Tf Nanoparticles

Anatase TiO₂ (10 mg) NPs were suspended in deionized water (1 mL) to prepare a working stock solution (10 mg/mL). 0.1 M of GdCl₃ stock solution was prepared by dissolving 1 mmol GdCl₃ (371.7 mg) in deionized water (10 mL). To load gadolinium onto TiO₂ nanoparticles, 250 μL TiO₂ stock solution was added to 30 μL GdCl₃ of the stock solution in 470 μL 1M, HEPES buffer (pH 7.1). Subsequently, the reaction mixture was incubated vigorously at 100°C for 1 h on an Eppendorf Thermomixer. Working stock solutions of Tf were also prepared by dissolving 30 mg of Tf in 1 mL PBS. To prepare TiO₂-Gd-Tf, a 1:3 (w/w) solution of TiO₂-Gd and Tf was mixed and sonicated in continuous mode for 2 min. It is essential to ensure the temperature of the solution does not exceed 55 °C to prevent denaturation of Tf (55°C) [6]. The NPs were then filtered through a 0.45 μm syringe filter to isolate monodispersed nanoparticles.

2.3. Size Characterization of TiO₂-Gd-Tf NPs

TEM images of the NPs were captured using a JEM-1400 series 120 kV Transmission Electron Microscope (JEOL USA, Inc, Peabody, MA, USA). Dynamic light scattering (DLS) measurements were acquired with a Malvern Zetasizer Nano ZS (Malvern Instruments Ltd., Malvern, UK) instrument equipped with a 633 nm laser. Three measurements were conducted for each sample with at least 10 runs. All sizes reported were based on the intensity average.

2.4. Colorimetric Assay for Gd Detection

Absorption spectrum of the samples was acquired using Beckman Coulter DU640 UV-Vis spectrophotometer (Beckman Coulter Inc, Brea, CA). Firstly, GdCl₃ samples with Gd concentration from 0 to 15 μM with an interval of 3 μM were prepared by mixing the appropriate amount of GdCl₃ stock solution with 10 μL Arsenazo III and diluted with DI water to a total volume of 1 mL. Similar experiments were performed in PBS buffer, as well as in different pH (2.98, 5.02, 6.91, 9.07, and 11.05), adjusted with hydrogen chloride or sodium hydroxide solution. pH values were acquired using Accumet AB150 pH meter (Thermo Fisher Scientific, Berkeley, MO, USA).

2.5. Inductively Coupled Plasma Mass Spectrometry (ICP-MS)

Determination of the NPs composition was performed using ICPMS. Briefly, 100 μL of samples were added to Teflon digestion vessels along with 2 mL of concentrated nitric acid (70%, trace metal grade). The sealed digestion vessels were then placed into a CEM Mars 6 microwave digestion system (CEM Co., Matthews, NC, USA). Microwave power was ramped to reach 200°C in 20 min; then that temperature was maintained for 20 min. The sample was then cooled and diluted to 3 mL with 18 MΩ DI water. A Nex-ION 2000 inductively coupled plasma mass spectrometer (PerkinElmer Inc, Waltham, MA, USA) was used to determine the Gd and Ti concentration of the diluted samples. Gd and Ti calibration standards of 1, 10, 50, 100, and 200 μg/L were used. The internal standard (200 μg/L, Sc, Tb) was continuously introduced during ICP-MS analysis. Ti and Gd concentration in the samples was calculated by multiplying the measured Ti, Gd concentration by the dilution factor.

2.6. ROS Generation Assay

DCF-DA was used to quantify general ROS production according to previous literature [14]. Briefly, DCF·DA was activated to DCF by adding 5 μL 1 N NaOH to DCF·DA (45 μL, 5.55 mM) in dimethyl sulfoxide (DMSO) and rested in a dark environment for 10 min, producing a 5 mM stock that was refreshed for each sample run. Activated DCF·DA (4 μL) was added to 0.01 mg/mL TiO₂-Gd-Tf samples to a total volume of 1 mL. An uncoated, black walled, and clear bottom 96 well plate (Greiner Bio-One) containing 150 μL sample per well was used for ROS quantification. Samples were performed in triplicate with a well geometry that allowed an average power of 1.9 mW across the plate. For comparison between runs, a bare 25 nm TiO₂ control was always plated to quantify variability. After loading, the plate was analyzed on a BioTek Synergy Neo2 plate reader (BioTek, Winooski, VT, USA) using 487 nm excitation and 528 nm emission. Subsequently, the plate was automatically exposed to UV light for 80 seconds before reading again, repeating the process. This was carried out for a total of 30 min for each plate and the data was compiled into pseudo-first-order kinetic curves for reporting.

2.7. Cell Culture

HT1080 cell line was purchased from American Type Culture Collection (ATCC, Manassas, VA, USA) that underwent STR profiling and were tested for mycoplasma contamination. HT1080 cells were cultured in Dulbecco’s Modified Eagle’s Medium containing 10% fetal bovine serum (FBS), L-glutamine (2 mM), penicillin (100 units/mL), and streptomycin (100 μg/mL), incubated at 37°C in a humidified atmosphere of 5% CO₂ and 95% air.

2.8. In Vitro Cell Viability Assays

Cell viability study was performed using MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assay, a colorimetric assay for assessing the viability of cell culture. The assay was performed using CellTiter 96 AQ_ueous Non-Radioactive Cell Proliferation Assay kit (Promega Co., Fitchburg, WI, USA) according to the manufacturer’s instructions. HT1080 cells were placed in a 24-well plate (TPP, Midwest Scientific, St. Louis, MO, USA) containing 400,000 cells per well, and were then incubated with TiO2-Gd-Tf NPs in the range of 25 to 150 μg/mL. Analysis was made after 48 h incubation.

2.9. In Vitro RaST Assay

In vitro RaST assay was performed similarly. HT1080 cells were placed in 24-well plates containing 400,000 cells per well. TiO₂-Gd-Tf NPs were added to the well with the final concentration of 100 μg/mL and incubated overnight for cell internalization. ¹⁸FDG (100 μCi/mL) was then added to each well and further incubated for 48 h. Cell viability was measured with CellTiter 96 AQ_ueous Non-Radioactive Cell Proliferation Assay kit according to the manufacturer’s instructions. Similar experiments used as reference controls were performed on cells treated with NPs alone or FDG alone.

2.10. Cell Internalization Study

AlexaTf was used to prepare fluorescent TiO₂-Gd-AlexaTf NPs, and the products were processed as described above. In general, HT1080 cells were cultured and incubated with TiO₂-Gd-AlexaTf with a final concentration of 50 μg/mL in Lab-Tek eight-chambered slides for 1 h, 4 h and 24 h at 37°C. After the treatment, the cells were washed by PBS for three times, live cell nucleus was stained by the culture medium containing SY-13 nuclear dye (Invitrogen, Carlsbad, CA) for 45 min at 37°C. Cells were mounted and covered with coverslip. Olympus FV1000 confocal laser scanning microscope was applied to visualize all samples using Ex/Em 488/500-530 nm for the nucleus and Ex/Em 674/700-760 nm for the NPs. Both FV1000 confocal and ImageJ software (National Institutes of Health) were used for image processing.

2.11. Relaxivity Measurement and MR Phantom Study

Relaxivity was firstly measured with 1.4 T Pulsar NMR Spectrometer (Oxford Instruments, Oxfordshire, UK). For each concentration (0, 0.02, 0.04, 0.06. 0.08 mM) of TiO₂-Gd-Tf NPs and Gd-DTPA, the samples were loaded into NMR tubes. The in vitro T1 and T2 were measured by using inversion recovery and a CPMG sequence, respectively.

The MRI properties of TiO₂-Gd-Tf NPs were evaluated and compared to Gd-DTPA by loading a series of concentrations of each into one row of an 8-well plastic plate. MRI was performed on a 4.7 T Agilent (Agilent Technologies, Santa Clara, CA, USA) DirectDrive small animal scanner using a 7-cm ID quadrature birdcage coil. For all sequences, a single 1 mm thick slice was acquired with a 6.4 x 3.2 cm field of view. T1-weighted and T2-weighted images were acquired at 0.25 x 0.25 mm resolution, while the T1 and T2 maps were acquired at 0.5 x 0.5 mm resolution.

Sequence-specific MRI parameters were as follows: T1-weighted spin echo image: TR/TE = 0.25/0.0135 seconds; T2-weighted fast spin echo image: TR/TE (effective) = 4/0.06 seconds, echo train length = 8; Inversion-recovery fast-spin echo images for T1 mapping: TR/TE = 10/0.057 seconds, effective echo time = 0.023 seconds, 8 exponentially spaced inversion times from 0.006 - 4 seconds; multi-echo spin echo images for T2 mapping: TR/TE = 8/0.01 seconds, 50 echoes. R1 and R2 maps were generated for each imaging voxel by modeling as a mono-exponential decay using the Bayesian Analysis Toolbox (bayesiananalysis.wustl.edu) [15, 16].

2.12. In Vivo Tumor Model

All studies were conducted in compliance with Washington University Animal Welfare Committee’s requirements for the care and use of laboratory animals in research. Athymic nu/nu nude mice (8-week-old) were purchased from Frederick Cancer Research and Development Center (Frederick, MD, USA). The HT1080 xenografts were generated by subcutaneous injection of 4 × 10⁶ cells in PBS (100 μL) on the back of Athymic nude mice.

2.13. In Vivo MR Imaging Studies

MRI scans were firstly performed on a 3 T Siemens Prisma scanner (Siemens Healthcare, Malvern, PA). The mouse was placed prone inside a knee coil after anesthesia. After scout imaging for localization, a free-breathing 3-dimensional (3D) fast low-angle shot (FLASH) sequence was used to obtain the volumetric images of the mouse. The scan was along the coronal direction with the following image parameters: TR/TE = 0.00505/0.00195 seconds, flip angle = 15°, bandwidth = 400 Hz/pixel, FOV = 82 x 82 x 26 mm³, matrix size = 484 x 484 x 44, interpolated voxel size = 0.17 mm x 0.17 mm x 0.6 mm, average number = 10, scan time = 9 minutes 50 seconds. The same 3D scans were performed pre-contrast, 8 hours, and 24 hours post-contrast. All 3D data sets were displayed as maximum intensity projection (MIP) images for further evaluation.

MRI scan was then performed on a 4.7 T small animal scanner using a 5-cm ID quadrature birdcage coil. 36 slices with 0.5 mm thickness were acquired with a 6.4 x 3.2 cm field of view, matrix size = 256 x 128, acquired resolution = 0.025 x 0.025 x 0.5 mm³. The following parameters were used for the scan: respiratory-gated fat-saturated T1-weighted 2D spin echo: TR/TE = 0.102/0.0123 seconds, 1 average, bandwidth = 195 Hz/pixel. The same scans were performed pre-contrast, 6 hours, and 24 hours post-contrast.

3. RESULTS AND DISCUSSION

3.1. Preparation of Gd-doped and Transferrin-coated TiO₂ NPs

Exchange of free water with Gd³⁺ ion hydration shells is critical for MRI contrast generation, requiring that Gd-doped metallic NPs do not overly impede water diffusion [17]. A method to accomplish this design goal is to attach Gd chelates to NPs. Previous studies have shown that this approach can enhance the relaxivity of Gd NPs, but dislodgment of the chelates from metallic nanoparticles creates discordance between MRI signal and the core NPs. Given our interest in tracking the distribution of core TiO₂ NPs for RaST, we used a chelate-free heat-induced method to functionalize TiO₂ NPs with Gd (Fig. 1A) [18]. Unfortunately, these particles were prone to aggregation in an aqueous medium. To stabilize the Gd on TiO₂ surface and minimize aggregation, we next coated the TiO₂-Gd NPs with Tf using a literature method to obtain TiO₂-Gd-Tf NPs [19]. At high concentrations of Tf, sonication facilitated the adsorption of Tf onto Gd-TiO₂ in neutral pH, which stabilized the NPs through negatively charged (isoelectric point = 5.5) protein-protein electrostatic repulsion [6]. An additional benefit of using Tf is that its receptors are overexpressed on the cell surface of many tumor types because of the high demand for iron by rapidly proliferating cells [20], doubling its potential use also to enhance tumor targeting.

Fig. (1).

Preparation and characterization of TiO₂-Gd-Tf NPs. (A) Schematic illustration of TiO₂-Gd-Tf NPs preparation. (B) Size distribution of TiO₂-Gd-Tf NPs from DLS. (C) Representative TEM image of TiO₂-Gd-Tf NPs, scale bar = 100 nm.

These TiO₂-Gd-Tf NPs were characterized with DLS, which showed an average size of 115 nm (Fig. 1B). The polydispersity index (PDI) of 0.173 indicates good monodispersity. Analysis by TEM showed spherical NPs with small nanoclusters of about 100 nm (Fig. 1C), suggesting that the DLS data represent defined clusters of NPs stabilized by Tf.

3.2. Simple Colorimetric Assay is Capable of Determining the Stability and Loading of Gd on TiO₂ NPs Compared to ICP-MS

ICP-MS is an established method to determine the composition of inorganic NPs, but the complexity of the procedure and the expensive nature of the method confine this analysis to specialized research centers. Previous studies have shown that Arsenazo III dye can be used to detect free Gd in aqueous solutions [21], presenting an opportunity to assess if this simple assay could be used to determine the loading and monitor the stability of Gd on TiO₂ NPs. To test the feasibility of this approach, we first determined the linearity of Gd detection in aqueous solutions and used the data to determine the stability and estimate the NPs loading capacity.

Spectral analysis of Arsenazo III in water at varying concentrations of Gd showed the expected red shift of the absorption spectrum from 537 nm to 652 nm peak upon Gd chelation (Fig. 2A). A proportional increase in the 652 nm peak correlated with increasing concentration of Gd, leading to a linear plot that can be used to quantify free Gd in solutions by colorimetric assay (Fig. 2B). Interestingly, attempts to determine free Gd in PBS by this colorimetric method were not successful as there was no enhancement of the 652 nm peak over a range of Gd concentrations (Fig. 2B). Given that Gd has a high affinity for phosphates [22], we attribute this loss of sensitivity to the chelation of Gd with the abundant phosphate group in PBS, which decreases the amount of free Gd below the limit of detection by Arsenazo III assay. Consequently, subsequent measurements were made in water.

Fig. (2).

Determination of TiO₂-Gd-Tf NPs stability and loading. (A) Absorption spectrum of a mixture of Arsenazo III with different concentrations of Gd. (B) Calibration curve of absorbance of the mixture at 652 nm against Gd concentration in water and PBS. (C) Calibration curves at different pH. (_D) Stability study of TiO₂-Gd-Tf NPs in water. The formulated NPs were incubated in water for 0, 1, 2, 4, 8 and 24 h. The NPs were centrifuged down and the supernatant was tested with the colorimetric assay, while the sediment was characterized with ICP-MS.

Stability of metal chelates depends on multiple factors, including the pH of the aqueous medium. To determine how pH affects the detection sensitivity of Gd, we measured the response of the 652 nm absorption in the pH range of ~3-11 (Fig. 2C). Irrespective of the pH, a linear correlation between Gd concentration and increase in the 652 nm peak was observed. However, differences in the slopes of the curves point to enhanced sensitivity of the assay below pH 6. This observation is probably consistent with the existence of Gd in predominantly Gd³⁺ species below pH 6 [22], which can favorably bind with the arsenous group (pKa2 = 1.2; pK3 = 2.7) even under low pH conditions. This result demonstrates that the chelation of Arsenazo III with Gd is pH dependent, and the lower the pH, the more sensitive for Arsenazo III to detect free Gd. The significant decrease in detection sensitivity at high pH of 11 suggests the formation of Gd(OH)₃, which prevents free Gd³⁺ from binding to Arsenazo III. Thus, controlling the pH of Arsenazo III solutions for determining free Gd will improve the detection sensitivity of the colorimetric assay.

Having established the optimal conditions for using Arsenazo III colorimetric assay to determine free Gd, we next explored its use to assess the stability of the TiO₂-Gd NPs. Given the amount of Gd used in the heat-induced doping reaction, we used the colorimetric assay to determine the amount of loaded Gd. After Tf coating, we used the same assay to monitor the release profile of Gd over time. Our result showed that the NPs were stable in an aqueous medium at 37°C up to the experimental time point of 24 h (Fig. 2D). The payload of Gd is 12% (Gd/TiO₂, w/w). ICP-MS is typically used as the gold standard for determining the compositions of metallic nanoparticles because of its high sensitivity [23]. Using the standard procedure [24], we performed ICP-MS analysis with the same batch of NPs used in the colorimetric assay under similar conditions. Comparison of both analysis methods demonstrated similar stability and loading of the Gd on TiO₂ NPs. The consistency between the two methods suggests that the simple and cost-effective colorimetric assay is sufficient for characterizing Gd-doped NPs.

3.3. TiO₂ NPs Retain their ROS-generating Capacity after Doping and Coating with Gd and Tf

Doping of TiO₂ NPs with metals such as Gd is known to alter its ROS-generating capacity due to changes in the electron hole recombination [25]. Given that RaST requires ROS production upon stimulation of TiO₂ to exert selective cytotoxicity, we used DCF·DA to quantify general ROS production, while PBS served as a reference [14]. Doping TiO₂ NPs with Gd significantly decreased ROS generation, probably due to the insertion of Gd ions into the defects of TiO₂ anatase (Fig. 3A). In contrast, subsequent coating with Tf recovered some of the lost photocatalytic activity. Given the ability of transferrin to form metallo-transferrin complexes [26], it is conceivable that some of the adsorbed Gd on TiO₂ NPs surface defects were trans-chelated by Tf, thereby freeing the electron recombination process and preventing Gd release into solution. This result supports the use of TiO₂-Gd-Tf NPs as a photosensitizer for RaST.

Fig. (3).

Photosensitizing and stimulated cytotoxic effects of TiO₂-Gd-Tf NPs. (A) ROS generation assay with various TiO₂ nanoconstructs and PBS buffer. DCF·DA was used to quantify ROS generation with the concentration of each NPs set to 0.01 mg/mL. (B) Cell toxicity of TiO₂-Gd-Tf using HT1080 cells. MTS assay was used after treating HT1080 cells for 48 h with the NPs concentrations ranging from 10 μg/mL to 150 μg/mL. PBS was used as a reference control. (C) Representative images of TiO₂-Gd-Tf internalization in HT1080 cells at 1 h, 4 h and 24 h after incubation. TiO₂-Gd NPs were coated with fluorescent AlexaTf. Blue indicates SYTO 13 nuclear stain and red indicates TiO₂-Gd-AlexaTf, scale bar = 20 μm. (D) In vitro RaST and controls. HT1080 cells were treated with the NPs only (100 μg/mL), FDG only (100 μCi/mL), and a combination of both NPs and FDG at the same NPs concentration and FDG activity.

Inorganic NPs can have pharmacological effects on cells due to the release of toxic metals or the induction of uncontrolled detrimental redox processes. With the goal of using the TiO₂-Gd-Tf NPs in living organisms, we evaluated their intrinsic toxic effects on cells using MTS colorimetric assay and the human sarcoma cells, HT1080. We found that the cells remained viable compared to untreated control up to 100 μg/mL (Fig. 3B). Above this concentration, a significant increase in cytotoxicity was observed, indicating that a wide range of TiO₂-Gd-Tf concentrations is suitable for induced toxicity by RaST.

An important requirement for RaST is the internalization of the nanophotosensitizers in cells. To visualize the cell uptake by microscopy, we coated TiO₂-Gd NPs with AlexaTf. Using the nontoxic concentration of TiO₂-Gd-AlexaTf NPs (50 μg/mL) and HT1080 cells, we observed a time-dependent uptake of the NPs in cells (Fig. 3C). While the intracellular distribution pattern was diffused, they remained perinuclear up to 24 h post-incubation. The diffuse profile of the NPs in HT1080 makes them more accessible to radiopharmaceuticals, which will serve as a source of Cerenkov radiation and stimulate ROS production for RaST. To test this hypothesis, we used ¹⁸FDG as our Cerenkov radiation source, as reported previously because of its ability to internalize in metabolically active cancer cells [27]. The amounts of ¹⁸FDG (100 μCi/mL) and TiO₂-Gd-Tf NPs (100 μ g/mL) were selected to ensure that they do not induce cytotoxic effects alone compared to the untreated control (Fig. 3D). As we postulated, cells treated with both TiO₂-Gd-Tf NPs and ¹⁸FDG exerted a sustainable cell killing effect, indicating that doping TiO₂ NPs with Gd retained its photosensitizing effect and response to radionuclide stimulation.

3.4. Enhancement of TiO₂-Gd-Tf NPs Relaxivity Depends on Magnetic Field Strength

The motivation for this study is to develop Gd contrast-mediated MRI-guided RaST. Contrast generation dictates that the doping process ensures access of Gd to water, with a preference for compositions that will improve the r1 or r2 relaxivity per Gd over clinically used agents such as Magnevist. Furthermore, clinical MRI scanners utilize a variety of magnetic field strengths, which exhibit different relaxivity values for the same NPs. Finally, coating of TiO₂-Gd with transferrin could also shield Gd from water exchange. Therefore, we evaluated the r1 and r2 of TiO₂-Gd-Tf NPs benchmarked against Gd-DTPA on standard 1.4 T NMR. The calculated r1 of TiO₂-Gd-Tf NPs (10.43 mM⁻¹s⁻¹) is much higher than that of Gd-DTPA (3.77 mM⁻¹s−¹; Fig. 4A). Similarly, the r2 of TiO₂-Gd-Tf NPs (13.34 mM⁻¹s⁻¹) is about twice that of Gd-DTPA (5.51 mM⁻¹s⁻¹) on a 1.4 T field relaxometer (Fig. 4B). These data are consistent with previous studies that obtained similar results using plain TiO₂-Gd [12]. Our study demonstrates that coating TiO₂-Gd NPs with Tf does not disrupt the MRI contrast generating capability of the NPs. We next performed a similar study using a 4.7-Tesla small animal MR scanner. T1- and T2-weighted imaging was obtained from an 8-well plate phantom containing Gd-DTPA and TiO₂-Gd-Tf NPs with concentrations ranging from 0.0375 mM to 0.3 mM. Visual inspection of the images shows that TiO₂-Gd-Tf NPs enhanced MRI signal compared with Gd-DTPA in T1-weighted imaging (Fig. 4C). A remarkable signal decay was observed in the T2-weighted images of TiO₂-Gd-Tf NPs relative to Gd-DTPA (Fig. 4D). To quantify the r1 and r2 of both Gd-DTPA and the NPs under similar conditions, T1- and T2-mapping was then performed. While the r1 of TiO₂-Gd-Tf NPs and Gd-DTPA is 8.13 mM⁻¹s⁻¹ and 4.89 mM⁻¹s⁻¹, respectively (Fig. 4E), the r2 of TiO₂-Gd-Tf NPs and Gd-DTPA is 87.15 mM⁻¹s⁻¹ and 3.85 mM⁻¹s⁻¹, respectively (Fig. 4F). These data demonstrate that the TiO₂-Gd-Tf NPs provide consistent high relaxivity at both low and high magnetic field strengths. The exceptionally high r2 at 4.7 T, which is greater than r1 by a factor of about 11, suggests that the Gd-doped NPs could serve as an efficient T2-weighted contrast agent.

Fig. (4).

MR enhancement effects of TiO₂-Gd-Tf NPs and Gd-DTPA at different magnetic field strengths. (A) T1 and (B) T2 relaxivity coefficient of TiO₂-Gd-Tf NPs and free Gd-DTPA on 1.4 T NMR. (C) T1-weighted and (D) T2-weighted images of TiO₂-Gd-Tf NPs (upper) and Gd-DTPA (lower) on 4.7 T MRI, and the concentration of the sample was 0.0375 mM, 0.075 mM, 0.15 mM, 0.3 mM (left to right). (E) T1 and (F) T2 relaxivity coefficients of TiO₂-Gd-Tf NPs and Gd-DTPA on 4.7 T MRI.

3.5. In Vivo MRI Demonstrates the Uptake of TiO₂-Gd-Tf NPs in Tumors at Different Magnetic Field Strengths

Encouraged by the phantom study result, in vivo T1-weighted MR imaging was performed on HT1080 tumor-bearing nude mice using 3-Tesla Siemens Prisma human MR scanner. Contrast enhancement in the tumor was not observed pre-injection of TiO₂-Gd-Tf NPs (Fig. 5A). By 8 h post-injection, high accumulation in tumor tissue and the abdominal area being evident. While MRI signal in the abdominal region decreased significantly at 24 h post-injection, retention of the NPs in tumor tissue persisted during this time. The high signal in the abdominal cavity at 8 h post-injection is consistent with the known hepatobiliary clearance pathway of many NPs [28]. Conversely, the tumor uptake and retention could be attributed to a combination of enhanced permeability and retention effect and to some extent, due to the binding of Tf to the Tf receptor expressed on HT1080 cells.

Fig. (5).

In vivo MRI of HT1080 tumor-bearing nude mice with TiO2-Gd-Tf NPs. The mice were given 100 μL of 1 mg/mL TiO₂-Gd-Tf NPs through tail vein injection. The red circle indicates the position of the tumor. (A) Representative T1-weighted images in 3 T human MR scanner at pre-injection; and 8 h and 24 h post-injection. (B) T1-weighted images in 4.7 T small animal MR scanner at pre-injection; and 6 h and 24 h post-injection.

To access the effects of different magnetic field strengths, we imaged the same HT1080 tumor model on a 4.7 T small animal scanner. Consistent with the data obtained from the 3 T scanner, low tumor contrast was observed in the pre-injection image (Fig. 5B). Similarly, signal enhancement in the tumor area at 6 h and 24 h post-intravenous injection of the NPs was observed, although the contrast was lower than what was observed at 3 T. Most likely, the high r2 of the NPs at higher magnetic fields counteracted the T1-weighted contrast. The in vivo results demonstrate that TiO₂-Gd-Tf NPs can serve as a contrast agent for MRI, providing a pathway for future MRI-guided RaST.

CONCLUSION

We designed TiO₂-Gd-Tf NPs and demonstrated that Tf coating stabilized minimized particle aggregation. Loading and stability studies showed that a simple colorimetric Arsenazo III assay exhibited similar results as the complex ICP-MS method, simplifying the characterization technique. Despite the coating of the NPs with Tf, we found that the ROS-generating capacity of TiO₂ was retained in TiO₂-Gd-Tf NPs, indicating its potential use for RaST. HT1080 cells treated with TiO₂-Gd-Tf NPs remained viable up to 100 μg/mL, indicating that the materials are biocompatible. Our results showed that treatment of HT1080 cells with non-cytotoxic concentrations of the NPs and FDG activity responded to RaST compared to the untreated and non-RaST controls, demonstrating the demonstrated HT1080 cell line responded well to ¹⁸FDG, suggesting the TiO₂-Gd-Tf NPs are excellent candidates for in vivo RaST. Measurement of the relaxivity showed that the r1 and r2 of TiO₂-Gd-Tf NPs depended on the magnetic field strength, with consistent higher r1 and r2 than Gd-DTPA. At 4.7 Tesla, the NPs became predominantly a T2-dominant MRI agent compared to lower fields. In vivo T1-weighted MRI on a 3 T human MIR scanner and a 4.7 T small animal MRI scanner showed the distribution of the NPs and highlighted the uptake in tumor tissue. While the tumor-targeting effect of the Tf was not studied here, it could combine with the enhanced permeability effect to facilitate NPs retention in the tumor [29]. These results support the feasibility of using TiO₂-Gd-Tf NPs for MRI-guided RaST.

LIST OF ABBREVIATIONS

AlexaTf

Transferrin, Alexa Fluor^™ 680 Conjugate

FDG

Fluorine-18 Deoxyglucose

Gd

Gadolinium

ICP-MS

Inductively Coupled Plasma Mass Spectrometry

MRI

Magnetic Resonance Imaging

NIR

Near-Infrared

NPs

Nanoparticles

PDT

Photodynamic Therapy

PET

Positron Emission Tomography

RaST

Radionuclid Stimulated Cancer Treatement

ROS

Reactive Oxygen Species

TEM

Transmission Electron Microscopy

Tf

Transferrin

UV

Ultrviolet