Multifunctional graphene oxide/iron oxide nanoparticles for magnetic targeted drug delivery dual magnetic resonance/fluorescence imaging and cancer sensing

Graphene Oxide (GO) has recently attracted substantial attention in biomedical field as an effective platform for biological sensing, tissue scaffolds and in vitro fluorescence imaging. However, the targeting modality and the capability of its in vivo detection have not been explored. To enhance the functionality of GO, we combine it with superparamagnetic iron oxide nanoparticles (Fe3O4 NPs) serving as a biocompatible magnetic drug delivery addends and magnetic resonance contrast agent for MRI. Synthesized GO-Fe3O4 conjugates have an average size of 260 nm and show low cytotoxicity comparable to that of GO. Fe3O4 nanoparticles provide superparamagnetic properties for magnetic targeted drug delivery allowing simple manipulation by the magnetic field and magnetic resonance imaging with high r2/r1 relaxivity ratios of ~10.7. GO-Fe3O4 retains pH-sensing capabilities of GO used in this work to detect cancer versus healthy environments in vitro and exhibits fluorescence in the visible for bioimaging. As a drug delivery platform GO-Fe3O4 shows successful fluorescence-tracked transport of hydrophobic doxorubicin non-covalently conjugated to GO with substantial loading and 2.5-fold improved efficacy. As a result, we propose GO-Fe3O4 nanoparticles as a novel multifunctional magnetic targeted platform for high efficacy drug delivery traced in vitro by GO fluorescence and in vivo via MRI capable of optical cancer detection.


Preparation of graphene oxide-iron oxide nanocomposite
5nm Fe 3 O 4 staring material was first activated for 4 hrs with APTES dissolved at 1% w/v in toluene. Activated Fe 3 O 4 NPs were washed with toluene to remove free APTES, sedimented via centrifugation and finally dispersed in water. APTES-functionalized iron oxide was further coupled with graphene oxide (GO). Graphene Oxide was dispersed in DI water at 450 μg/mL and ultrasonically treated for 30 to 60 min at 3W to decrease the size of GO flakes down to 250 nm for effective cell internalization. Treated GO and Fe 3 O 4 NPs were coupled in a conjugation reaction using 1mmol (EDC) and 1mmol of (NHS). After 6 hrs, samples were centrifugally washed with DI water three times to purify the product sedimented during the centrifugation. Acidity (pH 6.4) of GO suspension allows to run the coupling reaction with EDC in water without the presence of the buffer. It is reported that the reaction is less effective at higher pH, however in the pH range of 4.5 to 7.2 reaction was shown to take place. [51]. There are also some reports of this type of conjugation without buffer at pH 7. [52] Characterization Synthesized GO-Fe 3 O 4 conjugates were further characterized with Transmission Electron Microscopy (TEM JEOL JEM-2100) at 200 kV to assess the morphology, crystallinity and lattice spacing. Capacity of GO-Fe 3 O 4 as an MRI contrast agent was assessed via measuring relaxation times T 1 and T 2 with Bruker (Minispec mq60) Relaxometer at 1.41 T at 37˚C. This material took 18 seconds to bring all the material to the cube side by using a magnet. Fluorescence spectra of the nanoconjugates were measured with Horiba Scientific, SPEX NanoLog Spectrafluorometer with 400 nm excitation and the emission in the range of 420 to 762 nm. This emission was assessed at different pH conditions that were achieved by adding microliter aliquots of NaOH or HCl to yield pH in the range of 6 to 8.4.
Ferrozine assay was used to determine the iron concentration in this composite. In this assay 500 μL of aqueous GO-Fe 3 O 4 suspension was mixed with 500 μL of 12M HCl to dissolve Fe 3 O 4 NPs, 500 μL of 12M of NaOH to neutralize the solution and then with 120 uL of 2.8 M HONH 2 HCl in 4 M HCl, 50 μL of 10M Ammonium Acetate and 300 μL of 10 mM Ferrozine in 0.1M Ammonium Acetate to allow for the assessment of the iron content. Absorbance was measured at 562 nm using Agilent Technologies Cary 60 UV-Vis and compared with previously measured calibration curve constructed with FeCl 2 as a standard.

Doxorubicin complexation
Doxorubicin was complexed with GO-Fe 3 O 4 noncovalently at a concentration of 25.5 μg/mL by overnight coincubation with prior vortexing. Bound DOX-GO-Fe 3 O 4 nanocomposites were separated from uncomplexed drug with a magnetic field. Absorption spectra of the uncomplexed drug were used to find the concentration of that (16.5 μg/ml) and assess the percentage of the drug that got complexed representing loading efficiency. Starting with DOX concentration of 42 μg/ml therefore allows to load 25.5 μg/ml or 60.7% of the free DOX on GO-Fe 3 O 4 . Provided the stock GO-Fe 3 O 4 concentration of 127 μg/ml used for complexion and assessed via GO characteristic absorption, the loading of DOX onto GO-Fe 3 O 4 was calculated to reach 20 wt%.

Cellular uptake and imaging
In vitro imaging was performed in three different cell types: HEK-293 (Human embryonic kidney fibroblast), HeLa (Human cervical carcinoma) and MCF-7 (Human breast cancer). GO-Fe 3 O 4 or DOX-GO-Fe 3 O 4 formulations were introduced to cells at concentration of 15 μg/mL and analyzed at several time points ranging from 30 min to 27h. Internalization study was performed in HeLa cells washed PBS and fixed with 4% paraformaldehyde at 30 min, 1, 3, 12, 24-and 27-hours. For the pH-based detection of cancer versus healthy cells, HEK-293, MCF-7 and HeLa cells were treated with 15 μg/mL of GO-Fe 3 O 4 , the concentration was measured using freeze-drying and verified via absorption measurements. Although the conjugates themselves were not sterilized, however all other materials (solvents and glassware) used were sterile. The location of each formulation was assessed using the intrinsic GO-Fe 3 O 4 fluorescence emission in the visible. Images were taken with Olympus IX73 microscope coupled to photometrics camera PRIM 958. For internalization studies 480nm excitation and 535nm emission filters were used to selectively image GO-Fe 3 O 4 conjugates whereas for cancer detection study we utilized 480nm excitation for 535 nm green emission and 550nm excitation for 635 nm red emission. Location of GO inside the cells is not considered during the calculation of intracellular green/red ratios, rather, the signal from all inside the cell is accumulated. Over 100 cells were analyzed to yield the aforementioned green/red ratios representing pHsensing by the GO. Extracellular emission from GO was collected only from the samples that were not fixed and the medium was not replaced leaving all extracellular GO intact.

Cytotoxicity assays
MTT cytotoxicity assays were performed with 3 formulations: GO-Fe 3 O 4 , DOX-GO-Fe 3 O 4 and free DOX at the same concentrations of DOX derived from loading on DOX-GO-Fe 3 O 4 and up to 15 μg/mL-imaging concentration of GO-Fe 3 O 4 . This assay was used to detect metabolic activity in cells based in a colorimetric probe. MTT assay test is based on the ability of living cells to reduce tetrazole (yellow) to formazan (purple) with the mitochondrial reductase; cell survival rates care calculated based on the absorbance with the formazan formed. For each concentration tested in MTT assay we used four replicas to calculate the error bars. Two technical replicas were performed in each MTT assay. Two cancer cell lines (HeLa-Human cervical carcinoma, and MCF-7 -Human breast cancer) were used in this work, as well as one noncancer cell line (HEK-293, Human embryonic kidney fibroblast). Cells were obtained from ATCC and maintained in a Thermo-Scientific Midi 40 CO2 Incubator at 37.1˚C with 5% carbon dioxide, 95% air.

Structural characterization
The formation of GO-Fe 3 O 4 hybrids is achieved by a straightforward coupling reaction between superparamagnetic APTES-Fe 3 O 4 NPs and GO in the presence of coupling reagents EDC and NHS (Fig 1) Prior to coupling, GO flakes are ultrasonically processed to reduce flake size for effective cellular internalization.
[22] After 30 minutes of ultrasonic treatment GO flakes are reduced from micron-sized structures to an average size of 569 ± 310 nm (S1 Fig) and after 1 hour of treatment-to 257 ± 120 nm ( Fig 2B). Although to fully predict the capability of intracellular transport the charge and hydrophobicity of GO need to be taken in account, the smaller~250 nm nanoparticle sizes are expected to be more suitable for cellular internalization. [53][54][55]  In this work we evaluate the capacity of synthesized GO-Fe 3 O 4 conjugates for biomedical applications. We explore their ability to be manipulated by magnetic field for magnetic targeted therapy, their role as MRI contrast agents, fluorescence imaging capacity, the capability of cancer detection via optical pH-sensing and anticancer drug delivery.

Magnetic targeting and MRI contrast agent capabilities
As synthesized hybrids show pronounced magnetic behavior and can be manipulated in suspension via a regular magnet (Fig 2D). This indicates a potential for magnetic targeting to the organs that require increased uptake of the delivered therapeutic. Although magnetic delivery to animal models was not explored, this is likely to take place due to high responsiveness of the nanoconjugates to the magnetic field (shown in Fig 2D), and may be object of future investigations. For the specific application of magnetic resonance imaging, the quality of an MRI contrast agent is more precisely evaluated by the relaxivity parameters r 1 or r 2 , which describe the ability of a contrast agent to shorten the T 1 or T 2 relaxation times of water, rather than by T 1 and T 2 themselves. Thus, for GO-Fe 3 O 4 conjugates we evaluate longitudinal r 1 and transverse r 2 relaxivity. These values are calculated through the dependence between the inverse proton relaxation times and the iron concentration: In   Graphene oxide/iron oxide for drug delivery, magnetic resonance/fluorescence imaging and cancer sensing being far above T B , which means that no magnetic remanence is present and thus the magnetization of the samples vanishes if the applied magnetic field is switched off. [64,65] Fluorescence imaging and pH-sensing GO fluorescence emission detected in red/near-IR for the starting material [66] (Fig 4A) has experienced a substantial spectral change upon functionalization with Fe 3 O 4 showing a narrower feature centered at 500nm with a broad shoulder in the red/near-IR. Notably the emission intensity was not affected by the functionalization still suggesting GO-Fe 3 O 4 conjugates as effective candidates for in vitro fluorescence imaging. The emission is stable over several weeks and does not exhibit photobleaching or aggregation-related broadening.
As well as GO, [67] GO-Fe 3 O 4 conjugates exhibit pH response in their emission. However, unlike GO, the increase in pH from 6 to 8 here results into quenching of the 500nm feature with subsequent slight enhancement in the red/near-IR shoulder and an isosbestic point at 600nm. This is indicative of the spectraphotometric titration behavior that in GO [67] was attributed to protonation/deprotonation of functional groups affecting electronic environments surrounding those. The ratios of green/red (500nm/650nm) GO-Fe 3 O 4 emission intensities are calculated to be unique for each pH (S1 Table) providing the capability of pH-sensing on the nanoscale via an optical non-destructive method. This is highly applicable to cancer detection as cancerous environments are expected to have lower pH due to overexcretion of lactic acid by several cancer cell types. [68] In vitro imaging and cancer detection GO-Fe 3 O 4 introduced to HeLa cells exhibits observable green (532 nm) emission, at 30 min, 1, 3, 12, 24 and 27 hours post transfection (Fig 5A). At each time point the emission intensity is significantly above the autofluorescence background in control samples and can be detected intracellularly. Extracellular GO-Fe 3 O 4 is removed by repeated replacement when the cells are fixed with paraformaldehyde. In order to assess the optimal internalization time, we analyze over 100 cells at each time point for average emission intensity per unit emissive area. Intracellular emission is maximized at 3h post transfection (Fig 5B) indicating the optimal internalization timeline with the following decline. As GO shows no appreciable degradation or emission quenching over these time periods in cellular media (S4 Fig), we attribute intracellular emission decrease (Fig 5B) to slow excretion of GO-Fe 3 O 4 conjugates over time down to 47% of the maximum in 27h. thus a high number of cells (100) per time point was used for internalization analysis.
We further utilize pH-dependence of GO-Fe 3 O 4 emission to assess its cancer detection capability for cancer (HeLa and MCF-7) versus healthy (HEK-293) cellular environments in vitro. In order to account for potential variation of pH in different cancer cell environments we use two types of cancer cells and integrate the emission intensities in over a 100 fluorescence images to calculate average emission intensity per unit area at two different wavelengths in green and red. We anticipate that the number of cancer cells producing lactic acid would affect the capability of pH sensing by GO due to accessibility of all GO flakes to the acidic environments. Thus, we analyze over hundreds of cells to average out the response from those that may not be in equivalent environments within the imaging areas. Additionally, for our imaging experiments we estimate the cell density of 625 cells/mm 2 that is within the standard cell density range used for in vitro work, [69,70] indicating that pH sensing can be conducted using GO-Fe 3 O 4 conjugates in regular in vitro experiments. Unlike in the internalization study, here we refrain from replacing the medium and focus mostly on extracellular emission of GO-Fe 3 O 4 due to more complex pH environments inside the cells often subject to intracellular pH buffering. For pH sensing cells are not fixed thus allowing for GO to be present extracellularly. GO-Fe 3 O 4 emission in red (635 nm) and green (535 nm) recorded in every cancer and healthy cell line with the spectrally-filtered microscopy imaging system providing characteristic green/red intensity ratios for pH assessment. These green/red emission intensity ratios show observable differences for cancer versus healthy cells (Fig 6A) which is confirmed by statistical measurements over the ensemble of cells (Fig 6B). Here higher ratios are observed for more acidic cancer cell environments as expected from the spectral dependence (Fig 4B). The very magnitudes of the intracellular emission-derived ratios can differ from the ones calculated from spectral pH behavior, since emission in microscopy images is recorded within the range of spectral filters. However, the general trend of higher green/red ratios for acidic environments of cancer cells prevails with 4-5 fold difference between cancer and healthy cells. Such Graphene oxide/iron oxide for drug delivery, magnetic resonance/fluorescence imaging and cancer sensing significant detection ratio suggests a promising potential of GO-Fe 3 O 4 as optical pH-sensors of cancerous environments.

Drug delivery
The primary purpose of dual fluorescence/MRI imaging and pH-sensing capabilities of GO-Fe 3 O 4 conjugates is to track anticancer drug delivery and image therapeutics in biological cells and tissues while allowing for concomitant cancer detection. We assess the drug transport properties of GO-Fe 3 O 4 via the delivery of Doxorubicin non-covalently attached to GO surface. Doxorubicin (DOX) is an established chemotherapeutic that due to poor water solubility [71] has a need for nanocarrier delivery. [72,73] Its hydrophobic structure advantageously allows DOX to complex non-covalently with several drug delivery vehicles including carbon nanotubes and polymeric micelles. [74][75][76] Non-covalent functionalization may facilitate improved drug release and is known to preserve optical/electronic properties of the nanocarrier essential for imaging. [74] To achieve non-covalent DOX loading on GO-Fe 3 O 4 , DOX is vortexed and incubated with GO-Fe 3 O 4 conjugates overnight with no additional agitation necessary. conjugates are then separated from an unbound DOX with a strong magnet. The absorption spectra of unbound DOX remaining in the solution is used to calculate the efficiency of DOX loading (% of free DOX loaded) on GO-Fe 3 O 4 and the loading capacity (weight percent of loaded DOX to GO-Fe 3 O 4 ). Such optical approach yields high loading efficiency of 61.42% and a loading capacity of 0.2 mg of DOX per 1 mg of GO-Fe 3 O 4 resulting in 20 wt% loading. Several works centered on DOX delivery by graphene oxide report lower or similar loading, however, show no improvement in DOX efficacy when complexed to GO. [77][78][79] Some can achieve substantially higher loading [80], however, do not report efficacy and in order to maintain that loading utilize GO flakes of larger sizes that may complicate cellular internalization. DOX-GO-Fe 3 O 4 conjugates utilized in the current study in addition to drug delivery also provide the capacities for cancer detection, imaging and MRI sensing which makes the DOX-GO-Fe 3 O 4 formulation more advantageous for theragnostic. In order to fully assess the efficacy of DOX-GO-Fe 3 O 4 complexes we investigate both cellular internalization and cell viability in the presence of DOX-GO-Fe 3 O 4 against DOX only control.
Introduced to HeLa cells DOX-GO-Fe 3 O 4 formulation shows effective internalization within the cytoplasm similarly to that of GO-Fe 3 O 4 carriers at the 3h time point (Fig 7B). Fluorescence emission from GO-Fe 3 O 4 platform does not exhibit significant changes due to non-covalent complexation, thus, we expect only negligible fluorescence contribution from DOX likely quenched by GO platform.
As compared to free DOX, DOX-GO-Fe 3 O 4 conjugates provide significantly higher efficacy at lower concentrations derived from cancer cell apoptotic response (Fig 7) evaluated using an MTT assay in HeLa cells. DOX-GO-Fe 3 O 4 offers 2.5-fold decrease in cell viability down to 37% with respect to a free drug at only~0.3 μg/mL dose of DOX and~2μg/mL concentration of GO-Fe 3 O 4. The GO-Fe 3 O 4 concentration used here is that of the whole platform. To achieve a similar response unbound DOX requires~8-fold higher concentrations. A higher toxicity exhibited by DOX when delivered by GO-Fe3O4 can be likely explained by the improved transport and internalization with the nanomaterial delivery vehicle that is generally known to enhance the efficacy of delivered therapeutics [81][82][83] GO-Fe 3 O 4 on its own exhibits only mild cytotoxicity, comparable to that of GO, which cannot account for the substantially enhanced therapeutic effect of the combined DOX-GO--Fe 3 O 4 formulation. DOX delivery and imaging so far did not incorporate magnetic targeting that in the tissues via targeted delivery approach expected to produce higher accumulation and, therefore, further improved efficacy. This response verifies the improved GO-Fe 3 O 4 -mediated intracellular transport. An advantage of substantial loading capacity also allows to select a broad treatment range with only a small dose of nanoparticles. Although implausible in the present in vitro work we intend to further utilize the magnetic targeting for significantly improved delivery and efficacy [84] in the further in vivo studies.

Conclusions
In this work we have successfully synthesized and tested the feasibility of multifunctional GO-Fe 3 O 4 conjugates with capabilities of dual magnetic resonance/fluorescence imaging, magnetic manipulation for targeting, optical pH sensing and drug delivery. These novel nanoparticles have an average size of 250nm suitable for cellular internalization and show comparable to GO low cytotoxicity at imaging concentrations of 15 μg/mL. The relaxation properties of GO-Fe 3 O 4 conjugates are comparable to existing free nanoparticle analogs, GO-Fe 3 O 4 conjugates have potential of as negative MRI contrast agents. GO-Fe 3 O 4 conjugates can be effectively manipulated by a magnet in suspension which allows for direct magnetic targeted accumulation in a specific therapeutic site. The GO surface contains a variety of functional groups for covalent attachment of molecular therapeutics or a substantial hydrophobic graphene platform for non-covalent functionalization with aromaticbased drugs with poor water solubility. In our work GO-Fe 3 O 4 conjugates show efficient intracellular delivery of non-covalently attached Doxorubicin with considerable drug loading and over 2.5-fold improvement in its efficacy over free drug at low concentrations. This in turn allows using 8 times lower dose of Doxorubicin to achieve the same therapeutic effect of~62% cancer cell death. The therapeutic delivery is tracked by the intrinsic green fluorescence of GO-Fe 3 O 4 complex that indicates efficient internalization at 3 hours post transfection with further excretion from the cells. The pH-dependence of this emission allows using the ratios of emission intensity in green (535 nm) to red (635 nm) to differentiate between cancer (MCF-7 and HeLa) and healthy (HEK-293) extracellular environments with a substantial 4 to 5-fold difference. As a result, we propose GO-Fe 3 O 4 as a unique multifunctional nanomaterial for magnetic-targeted drug delivery, dual in vitro fluorescence and in vivo MRI imaging and optical detection of cancerous environments.