MRI Detection and Therapeutic Enhancement of Ferumoxytol Internalization in Glioblastoma Cells

Recently, the FDA-approved iron oxide nanoparticle, ferumoxytol, has been found to enhance the efficacy of pharmacological ascorbate (AscH−) in treating glioblastoma, as AscH− reduces the Fe3+ sites in the nanoparticle core. Given the iron oxidation state specificity of T2* relaxation mapping, this study aims to investigate the ability of T2* relaxation to monitor the reduction of ferumoxytol by AscH− with respect to its in vitro therapeutic enhancement. This study employed an in vitro glioblastoma MRI model system to investigate the chemical interaction of ferumoxytol with T2* mapping. Lipofectamine was utilized to facilitate ferumoxytol internalization and assess intracellular versus extracellular chemistry. In vitro T2* mapping successfully detected an AscH−-mediated reduction of ferumoxytol (25.6 ms versus 2.8 ms for FMX alone). The T2* relaxation technique identified the release of Fe2+ from ferumoxytol by AscH− in glioblastoma cells. However, the high iron content of ferumoxytol limited T2* ability to differentiate between the external and internal reduction of ferumoxytol by AscH− (ΔT2* = +839% for external FMX and +1112% for internal FMX reduction). Notably, the internalization of ferumoxytol significantly enhances its ability to promote AscH− toxicity (dose enhancement ratio for extracellular FMX = 1.16 versus 1.54 for intracellular FMX). These data provide valuable insights into the MR-based nanotheranostic application of ferumoxytol and AscH− therapy for glioblastoma management. Future developmental efforts, such as FMX surface modifications, may be warranted to enhance this approach further.


Introduction
Ferumoxytol (Feraheme ® , FMX) is a clinically available, superparamagnetic iron oxide nanoparticle approved for treating iron deficiency anemia in patients with chronic kidney disease [1][2][3][4].FMX can generate T1-contrast enhancement in tumor tissue in glioma imaging due to its ferromagnetic properties.FMX is a superparamagnetic iron oxide nanoparticle (SPION) with a Fe 3 O 4 core that is about 30 nm in size, has a neutral charge, and resides within a carboxylated polymer coating [5].Many units of the Fe 3 O 4 core exist in one nanoparticle yielding a wide range of molecular weights with an average of about 730 kDa [6].Because of the large iron content of one molecule of FMX (1 molecule has ≈ 5900 iron atoms or 1 nM FMX ≈ 5.9 µM iron), it can function as a T 1 /T 2 * MRI contrast agent [7,8].Ferumoxytol's iron content and ferromagnetic properties also allow its use as a T 2 *-contrast agent because T 2 * relaxation times are largely influenced by paramagnetic and ferromagnetic materials (e.g., iron).FMX's superparamagnetic properties alter T 2 * relaxation times [9,10].FMX is an attractive MR contrast agent because it has a significantly longer intravascular half-life (t 1/2 ≈ 14-21 h) than gadolinium-based compounds (t 1/2 ≈ 1 h) [7,11].
Beyond its utility as an MRI contrast agent, FMX has shown potential as an anti-cancer therapy [12,13].The anti-cancer mechanism of FMX has been suggested to be due to its redox activity.It has previously been shown that the Fe 3 O 4 core can be oxidized by ionizing radiation, showing that FMX can serve as a reserve of redox-active iron [14].FMX also reacts with H 2 O 2 stimulating the release of iron from the nanoparticle.Thus, FMX may undergo redox reactions with a wide array of species.Ascorbate (AscH − ) is a one-electron reductant that can readily reduce some complexes of ferric (Fe 3+ ) to ferrous (Fe 2+ ) iron [15].AscH − can reduce and release Fe 2+ from ferritin, a Fe 3+ -containing biological macromolecule that is the primary mechanism for intracellular iron storage [16][17][18].Recently, it has been reported that AscH − catalyzes the decomposition of the FMX Fe 3 O 4 core [19].The chemical interaction between FMX and AscH − can be characterized by a significant reduction in FMX size (≈66% reduction in 24 h), a release of redox-active Fe 2+ that follows Michaelis-Menton kinetics, and a significant increase in H 2 O 2 generation.The decomposition of FMX by AscH − was reported to enhance glioblastoma cell killing and importantly, the enhanced toxicity of FMX and AscH − was glioblastoma specific, as no significant in vitro toxicity was observed in normal human astrocytes [19].Thus, the chemical pairing of FMX and AscH − represents a novel therapeutic strategy.However, the utility of FMX as an MRI contrast agent suggests that FMX and AscH − may have nanotheranostic potential.
T 2 * relaxation mapping is a quantitative MRI technique used primarily to indicate total iron content [20].This technique is widely applicable clinically for cardiac and hepatic iron overload [21][22][23][24][25][26].Recent studies have shown that beyond total iron content, T 2 * can provide information on the oxidation state of iron, specificity differentiating between Fe 3+ and Fe 2+ [27,28].This effect is theorized to result from proton-electron dipoledipole interactions associated with the number of unpaired electrons (i.e., electron spin magnetic moment) of transition metals [29].Moreover, a recent phase 2 clinical trial testing AscH − therapy in combination with radiation and temozolomide showed that patients with short T 2 * relaxation times (i.e., high iron content) had significantly greater therapy responses [30].Because T 2 * relaxation appears to be largely dependent on the paramagnetic properties of metals and can detect alterations in these electronic spin properties (e.g., reduction/oxidation of iron), T 2 * mapping may serve as a useful tool in the evaluation of FMX redox chemistry.Therefore, changes in T 2 * relaxation may be reflective of the disruption of the FMX Fe 3 O 4 core by AscH − .This study aims to provide detailed proof-ofconcept insights into the ability of T 2 * mapping to evaluate Fe 3 O 4 disruptions by AscH − with respect to the in vitro toxicity of extracellular and intracellular FMX.

In Vitro MRI Studies
Glioblastoma cells were treated with 20 pmol cell −1 AscH − for 1 h with 20 µg mL −1 FMX or pre-incubated for 24 h with 20 µg mL −1 FMX-L prior to the 1 h AscH −1 treatment.Following treatment, cells were trypsinized, re-suspended in sterile PBS, and transferred to PCR wells embedded in a 1% agarose gel phantom.Cells were allowed to collect at the bottom of the PCR well to form a pellet to be imaged.Cell pellets were then imaged on a 7T GE MR901 small animal scanner, a part of the small animal imaging core at the University of Iowa.T 2 * weighted images were collected using a gradient-echo sequence (TR = 10 ms, TE = 2.2, 8.2, 14.2, and 20.2 ms, matrix = 256 × 256, FOV = 25 × 20 mm, 2 signal averages).A B 0 shimming routine was performed to limit the effect of macroscopic field inhomogeneities.T 2 * maps were generated using a combination of 4 echo times collected and fitting each voxel to a mono-exponential curve using in-house Python code.Images were imported to 3D Slicer software (V5.0.3)where regions of interest (ROIs) were delineated as a 1 mm diameter cylinder in the center of the tube and mean T 2 * values were calculated using the label statistics tool within 3D Slicer [31].

FMX Internalization with Lipofectamine
Lipofectamine FMX (FMX-L) was generated using the commercially available lipofectamine 3000 reagents (Thermofisher Scientific, Waltham, MA, USA; L3000015).Functionalization was completed by diluting FMX at 1:16 in 1% FBS containing DMEM-F12 media (1 mL) with 10 µL P3000 reagent, vortexing vigorously for 5 s, and then diluting the FMX/P3000 stock at 1:1 with lipofectamine 3000.The samples were incubated at room temperature for 15 min prior to utilization.FMX-L was generated new for every experiment.The cells were then treated with FMX-L for 24 h in 1% FBS containing DMEM-F12 medium.The cells were washed with 1X D-PBS prior to additional studies to remove extracellular FMX.

Quantitation of Intracellular Iron
Intracellular iron concentrations were validated colorimetrically following a 24 h treatment with either 20 µg mL −1 FMX or FMX-L using a ferrozine-based assay [32,33].Following treatment, cells were washed with sterile PBS, trypsinized, and centrifuged at 1200 rpm for 5 min.The cell pellets were resuspended in 1X RIPA buffer (Sigma-Aldrich, St. Louis, MO, USA; R0278) and sonicated 3 × 10 s to lyse the cells.Cell lysis solution was then diluted 1:1 in 2.5 M glacial acetic acid pH = 4.5 with 5 mM ferrozine and 10 mM AscH − .The sample and buffer mixture were centrifuged at maximum speed (14,000× g) for 10 min to remove protein aggregates.The supernatant (200 µL) was placed in a 96-well dish [33].Ultraviolet-visible light (UV-Vis) spectroscopic measurements were performed using a 96-well plate reader.Fe 2+ (ferrozine) 3 complex formation was monitored by analyzing absorbance at 562 nm.Fe 2+ concentrations were determined using Beer's Law for absorbance at 562 nm (ε 562 = 27,900 M −1 cm −1 ) with a path length, of L = 0.55 cm (200 µL sample).

Cellular Iron Staining
To visualize the iron deposition following FMX treatment, cells were stained using a Prussian Blue technique using an iron staining kit (Abcam, Cambridge, U.K.; ab150674) using the manufacturer's protocol.Following treatment, cells were washed with 1X D-PBS and fixed with formalin for 5 min.The cells were then washed with distilled H 2 O and incubated for 15 min with a 1:1 mixture of potassium ferrocyanide and 2% hydrochloric acid.After staining, cells were washed with distilled H 2 O and stained for 5 min with a nuclearfast red counterstain.Finally, cells were washed with distilled H 2 O and allowed to dry.The cells were then imaged using a phase contrast microscope with a 40× objective lens.

Electron Paramagnetic Resonance Evaluation of FMX Concentrations in Cell Culture Media
The FMX concentrations were determined by measuring the peak-to-peak signal intensity of the EPR spectrum of the low-spin Fe 3 O 4 complex at g ≈ 2 as previously described [14].Using a Bruker EMX spectrometer, the following scan parameters were used to collect spectra: center field = 3508.97G, sweep width = 2000 G, frequency = 9.85 GHz, power attenuation = 18 dB, modulation frequency = 100 kHz, modulation amplitude = 0.7 G, with spectra being generated from a signal average of 2 scans with 2048 resolution.U87 cells were incubated for 24 h with 20 µg mL −1 FMX or FMX-L.

In Vitro Oxidation State Specificity of T 2 * Mapping
Before evaluating if T 2 * mapping can detect FMX and AscH − chemistry, the in vitro oxidation state specificity of T 2 * mapping was tested using a previously established MRI phantom model system [29].It was observed that AscH − increased T 2 * relaxation times in U87, U251, and U118 GBM cell lines by 7 ms, 17 ms, and 10 ms, respectively (Figure 1).This is consistent with the previously observed increase in T 2 * relaxation times following a pharmacological ascorbate infusion in GBM subjects [27].Moreover, the iron chelator deferoxamine (DFO) causes a decrease (U87 = −12 ms, U251 = −6 ms, and U118 = −18 ms) in T 2 * relaxation times indicative of a paramagnetic shift as a result of ferrioxamine (DFO-Fe 3+ ) complex formation.This is consistent with the ability of DFO to bind and maintain Fe in the +3 oxidation state (Fe 3+ ) [34].Thus, T 2 * mapping can detect iron oxidation state changes associated with the oxidation when complexed by DFO or internally reduced by AscH − .spectra being generated from a signal average of 2 scans with 2048 resolution.U87 cells were incubated for 24 h with 20 µg mL −1 FMX or FMX-L.

In Vitro Oxidation State Specificity of T2* Mapping
Before evaluating if T2* mapping can detect FMX and AscH − chemistry, the in vitro oxidation state specificity of T2* mapping was tested using a previously established MRI phantom model system [29].It was observed that AscH − increased T2* relaxation times in U87, U251, and U118 GBM cell lines by 7 ms, 17 ms, and 10 ms, respectively (Figure 1).This is consistent with the previously observed increase in T2* relaxation times following a pharmacological ascorbate infusion in GBM subjects [27].Moreover, the iron chelator deferoxamine (DFO) causes a decrease (U87 = −12 ms, U251 = −6 ms, and U118 = −18 ms) in T2* relaxation times indicative of a paramagnetic shift as a result of ferrioxamine (DFO-Fe 3+ ) complex formation.This is consistent with the ability of DFO to bind and maintain Fe in the +3 oxidation state (Fe 3+ ) [34].Thus, T2* mapping can detect iron oxidation state changes associated with the oxidation when complexed by DFO or internally reduced by AscH − .

Lipofectamine Enhances FMX Internalization
A potential limitation of this approach is the extracellular nature of FMX [35].Therefore, a proof-of-concept internalization model using lipofectamine was used to determine if T2* mapping can distinguish intracellular and extracellular FMX reduction by AscH − .To validate this model system, U87 cells were incubated with 20 µg mL −1 FMX ± lipofectamine (FMX-L) for 24 h.The initial observation made using this approach was that cell pellets following treatment with FMX-L had a reddish hue that would be indicative of high iron content (Figure 2a).Quantitatively, there was a significant decrease in FMX concentrations in the cell culture media, evaluated using EPR spectroscopy (Figure 2b) [14].This indicates a shift of FMX from the extracellular to the intracellular space.The cell pellets also showed a significant, ≥3-fold, increase in iron concentrations (Figure 2c).This was further validated using Prussian blue staining where intracellular iron was markedly increased following FMX-L treatment (Figure 2d).Interestingly, an increase in Prussian blue positive cells was visible following a 1 h FMX incubation.This effect was not as pronounced by 24 h.This suggests an initial extracellular accumulation of FMX that dissipates over time.Lipofectamine appears to be a valuable tool for facilitating FMX internalization and intracellular retention.

Lipofectamine Enhances FMX Internalization
A potential limitation of this approach is the extracellular nature of FMX [35].Therefore, a proof-of-concept internalization model using lipofectamine was used to determine if T 2 * mapping can distinguish intracellular and extracellular FMX reduction by AscH − .To validate this model system, U87 cells were incubated with 20 µg mL −1 FMX ± lipofectamine (FMX-L) for 24 h.The initial observation made using this approach was that cell pellets following treatment with FMX-L had a reddish hue that would be indicative of high iron content (Figure 2a).Quantitatively, there was a significant decrease in FMX concentrations in the cell culture media, evaluated using EPR spectroscopy (Figure 2b) [14].This indicates a shift of FMX from the extracellular to the intracellular space.The cell pellets also showed a significant, ≥3-fold, increase in iron concentrations (Figure 2c).This was further validated using Prussian blue staining where intracellular iron was markedly increased following FMX-L treatment (Figure 2d).Interestingly, an increase in Prussian blue positive cells was visible following a 1 h FMX incubation.This effect was not as pronounced by 24 h.This suggests an initial extracellular accumulation of FMX that dissipates over time.Lipofectamine appears to be a valuable tool for facilitating FMX internalization and intracellular retention.

FMX Internalization Enhances AscH − Cytotoxicity
This FMX internalization model system was used to evaluate if changes in T2* relaxation times reflect the internal reduction of FMX by AscH − .U87 cells were either co-incubated for 1 h with 20 µg mL −1 FMX ± 20 pmol cell −1 AscH − or pre-treated for 24 h FMX-L to load the cells with FMX prior to their 1 h AscH − treatment.Following treatment, cells were pelleted for T2* map generation.From this experiment, it has been observed that

FMX Internalization Enhances AscH − Cytotoxicity
This FMX internalization model system was used to evaluate if changes in T 2 * relaxation times reflect the internal reduction of FMX by AscH − .U87 cells were either co-incubated for 1 h with 20 µg mL −1 FMX ± 20 pmol cell −1 AscH − or pre-treated for 24 h FMX-L to load the cells with FMX prior to their 1 h AscH − treatment.Following treatment, cells were pelleted for T 2 * map generation.From this experiment, it has been observed that following a 1 h treatment with FMX or a 24 h treatment with FMX-L caused a noticeable signal loss, likely due to the ferromagnetic properties of FMX (Figure 3a).In both FMX and FMX-L treated cells, there was an observable susceptibility artifact surrounding the cell pellet that was much larger in the FMX-L cells, indicative of the significant increases in intracellular iron content that were previously described.AscH − -treated cells showed longer T 2 * relaxation properties; however, this was difficult to qualitatively visualize in the FMX-L treated cells due to the large signal loss.Quantitatively, AscH − alone induced a 5 ms increase (control = 25.6 ms versus AscH − = 30.4ms) in T 2 relaxation time, consistent with previous reports (Figure 3b) [27].Both FMX and FMX-L cells caused a decrease in T 2 * relaxation time to 2.8 and 1.9 ms, respectively.This is consistent with the observed FMX deposition with both treatments.In both cases (FMX and FMX-L), AscH − treated cells had significantly longer T 2 * relaxation times (25.6 and 22.3 ms, respectively).The T2* relaxation time change from baseline was significantly greater in those cells treated with FMX/FMX-L and AscH − than AscH − alone (Figure 3c).However, the internalization of FMX only partially increased the change in T 2 * by AscH − , suggesting that these doses of FMX for extracellular/intracellular differentiation were likely in the signal saturation range.Overall, these results further support the hypothesis that T 2 * relaxation time can detect the reduction of FMX by AscH − , but the high iron content of FMX may limit this effect.

Discussion
This study describes the ability of T2* mapping to detect the release of ferrous iron from FMX by AscH − .The primary utilization of FMX in the context of glioblastoma management is as an MR contrast agent [7,36,37].FMX is also being investigated as a marker for glioblastoma progression [37].Therefore, T2* may also be a valuable tool to identify regions of FMX accumulation.We demonstrate that FMX can decrease T2* relaxation times in vitro.This is consistent with previous data showing that FMX can decrease T2* relaxation times in humans 24 h following its administration likely owing to its 14-21 h intravascular half-life [7,38].In this study, supraphysiological concentrations of AscH − (10 mM), which are typically achieved via intravenous injection during glioblastoma therapy, were Moreover, it has recently been reported that the combination of FMX and AscH − exhibited enhanced cytotoxic effects in glioblastoma cells and significantly enhanced the standard of care therapy (radiation and temozolomide) in an in vivo animal model [19].Thus, the therapeutic aspect of these imaging results was subsequently evaluated in glioblastoma cells.Based on the potential effects of FMX internalization on the ability of T 2 * to detect nanoparticle reduction, the effects on AscH − toxicity were evaluated.Consistent with these imaging results, FMX-L significantly enhanced the dose-dependent AscH − toxicity in U87 cells as FMX had a dose-enhancement ratio of 1.16 (p = 0.09) as compared to 1.54 for FMX-L (p < 0.05; Figure 3d).Thus, it appears that the internalization of FMX represents a novel strategy to enhance its utility in combination with AscH − ; however, this may be a context-dependent effect that warrants further consideration.

Discussion
This study describes the ability of T 2 * mapping to detect the release of ferrous iron from FMX by AscH − .The primary utilization of FMX in the context of glioblastoma management is as an MR contrast agent [7,36,37].FMX is also being investigated as a marker for glioblastoma progression [37].Therefore, T 2 * may also be a valuable tool to identify regions of FMX accumulation.We demonstrate that FMX can decrease T 2 * relaxation times in vitro.This is consistent with previous data showing that FMX can decrease T 2 * relaxation times in humans 24 h following its administration likely owing to its 14-21 h intravascular half-life [7,38].In this study, supraphysiological concentrations of AscH − (10 mM), which are typically achieved via intravenous injection during glioblastoma therapy, were used [39,40].Thus, this chemical combination more closely replicates an interaction that may be observed during glioblastoma therapy.Adding a reducing agent (AscH − ) to FMX increases T 2 * relaxation times, which coincides with the release of Fe 2+ from the nanoparticle core [19].This is consistent with the iron oxidation state specificity of T 2 * mapping [29].The oxidation state specificity of T 2 * mapping could be further validated in vitro in this study as AscH − induces an increase in T 2 * relaxation while DFO causes a decrease.Importantly, this chemistry effect was able to be replicated in the context of AscH − and FMX chemistry as the addition of AscH − can prolong FMX relaxation times.This indicates that AscH − can reduce the Fe 3+ sites of FMX leading to an increase in the Fe 2+ :Fe 3+ ratio, which can be detected with T 2 * mapping.These results are consistent with the increase in T 2 * associated with adding AscH − to FMX in an orthotopic glioblastoma model [19].Thus, the present study provides further insights into the ability of T 2 * mapping to detect the catalyzed release of Fe 2+ from the Fe 3 O 4 core by AscH − .FMX and AscH − chemistry was detected in both the extracellular and intracellular space with FMX internalization facilitated by lipofectamine.In this cell culture model, adding FMX caused a significant decrease in T 2 * regardless of its localization.The internalization did appear to shorten T2* relaxation times further, consistent with the significant increase in cellular iron content; however, detectable differences were challenging due to potential signal saturation.In both cases, FMX and FMX-L, adding AscH − significantly increased T 2 * relaxation times.Following the internalization of FMX (FMX-L), the increase in T 2 * relaxation time induced by AscH − was slightly greater but was ultimately limited by the potential signal saturation caused by FMX.Thus, it is important to note that due to the large size (≈30 nm) and high iron content of FMX, T 2 * relaxation appears to lose the ability to detect intracellular versus extracellular localization [19,41].Therefore, the use of T 2 * may have an intrinsic technical limitation where the high iron concentrations of FMX limit the range of oxidation state specificity and impair the ability to evaluate FMX reduction by AscH − .This can be overcome by using ultrashort echo time (UTE)-T 2 * and may warrant further investigation [42].
Furthering the nanotheranostic potential of FMX and AscH − , the internalization of FMX significantly enhanced AscH − toxicity.Thus, the internalization of FMX may significantly enhance the therapeutic utility in combination with AscH − in GBM.Developmental efforts have been previously put forth to functionalize FMX and enhance tumor traffick-ing and internalization.For example, it has been shown that FMX functionalized with a Toll-like receptor 3 agonist enhanced melanoma tumor control [43].Moreover, the trend towards a greater increase in T 2 * relaxation following internalization suggests that FMX reduction by AscH − is driving the enhanced toxicity.These results are also consistent with previous literature that demonstrates increases in intracellular iron content enhance AscH − toxicity [44].This would support the hypothesis that cellular AscH − uptake by sodium vitamin C transporters (SVCTs) mediate AscH − toxicity in glioblastoma cells [45].Therefore, it can be hypothesized that surface modifications of FMX to increase tumor trafficking and internalization can enhance the effectiveness of FMX and AscH − in the management of GBM and warrant further investigation.

Figure 2 .
Figure 2. T2* mapping detects FMX internalization and reduction in vitro.(a) Cells were treated for 24 h followed by PBS washing and trypsinization.The large increase in intracellular iron content of FMX-L becomes apparent due to the reddish hue of the cell pellet.(b) Relative [FMX] concentrations in cell culture media following 24 h incubation.This was done by evaluating the EPR spectral peak of FMX at t = 0 and t = 24 h and normalizing both FMX and FMX-L peaks to FMX alone.(c) Intracellular, chelatable iron content in U87 cells following a 24 h incubation with FMX or FMX-L.Error bars represent mean ± SEM with * p < 0.05 using a Welch's t-test.(d) Representative phase contrast (40×) Prussian blue images for cellular iron content in U87 cells treated with FMX for 1 h and 24 h, or 24 h FMX-L.Black arrows indicate clusters of Prussian blue-positive cells.

Figure 2 .
Figure 2. T 2 * mapping detects FMX internalization and reduction in vitro.(a) Cells were treated for 24 h followed by PBS washing and trypsinization.The large increase in intracellular iron content of FMX-L becomes apparent due to the reddish hue of the cell pellet.(b) Relative [FMX] concentrations in cell culture media following 24 h incubation.This was done by evaluating the EPR spectral peak of FMX at t = 0 and t = 24 h and normalizing both FMX and FMX-L peaks to FMX alone.(c) Intracellular, chelatable iron content in U87 cells following a 24 h incubation with FMX or FMX-L.Error bars represent mean ± SEM with * p < 0.05 using a Welch's t-test.(d) Representative phase contrast (40×) Prussian blue images for cellular iron content in U87 cells treated with FMX for 1 h and 24 h, or 24 h FMX-L.Black arrows indicate clusters of Prussian blue-positive cells.