Distributions of intravenous injected iodine nanoparticles in orthotopic u87 human glioma xenografts over time and tumor therapy

Abstract

Aim:

To analyze the localization, distribution and effect of iodine nanoparticles (INPs) on radiation therapy (RT) in advanced intracerebral gliomas over time after intravenous injection.

Materials & methods:

Luciferase/td-tomato expressing U87 human glioma cells were implanted into mice which were injected intravenously with INPs. Mice with gliomas were followed for tumor progression and survival. Immune-stained mouse brain sections were examined and quantified by confocal fluorescence microscopy.

Results:

INPs injected intravenously 3 days prior to RT, compared with 1 day, showed greater association with CD31-staining structures, accumulated inside tumor cells more, covered more of the tumor cell surface and trended toward increased median survival.

Conclusion:

INP persistence and redistribution in tumors over time may enable greater RT enhancement and clinically relevant hypo-fractionated-RT and may enhance INP efficacy.

Nearly 24,000 people in the USA are diagnosed with primary brain tumors every year; metastatic brain tumor diagnoses are approximately 200,000. By the time the primary tumors are discovered and diagnosed, tumor cells have migrated along blood vessels and white matter tracts over large regions of the brain thereby limiting effective therapy [1]. Brain metastases may seed multiple locations throughout the brain by the time they are discovered [2]. Highly effective radiation cannot be delivered to such large portions of the brain due to the damaging effects of radiation on the normal brain, especially in children [3] whose brains are still developing. Therefore, radiation therapy (RT) for brain tumor therapy is largely palliative. Gold nanoparticles (AuNPs), iodine nanoparticles (INPs) and other high atomic number nanoparticles or compounds have the potential to increase the specificity of RT to the tumor cells by enhancing the local dose to tumor regions preferentially [4,5]; by virtue of the photoelectric effect, electrons from K and L shells are released from the heavy atoms after x-irradiation and these electrons deposit their energy locally [6] increasing free radical generation in and around the tumor cells. Previous studies with intracerebral gliomas show that both AuNPs and INPs localize to the tumor region and even migrated tumor cells preferentially after intravenous (iv.) injection [7,8]; the resulting increase in RT dose to the tumor increases tumor damage and significantly extends median survival of mice with advanced orthotopic gliomas [8–10] and metastatic tumors (Hainfeld et al. [11]). Recently developed INPs do not manifest detectable toxicity, do not affect thyroid function and overcome skin discoloration and pharmacokinetic drawbacks associated with the AuNPs [12]; when injected iv. prior to RT, the INPs were shown to increase median survival of mice with advanced intracerebral U87 gliomas more than twofold when compared with RT alone [8]. INP-enhanced RT (INP-RT) was also shown to synergize with chemotherapy [8]. Increased survival of mice with advanced intracerebral breast-derived tumors was even more pronounced with significant numbers of mice surviving greater than five-times the median survival of mice treated with only RT (Hainfeld et al. [11]). The mechanisms by which the INPs prolong survival are not fully known and need further study, and likely include increased radiation damage to tumor DNA, for example, double strand breaks (Smilowitz et al., unpublished observations) and increased endothelial cell damage; previously we showed that iv. injected INPs localize preferentially to CD31 containing structures largely in and around U87 tumors [8]. In this study, both the localization and distribution of the INPs to regions of advanced U87 tumors and efficacy were compared 1 and 3 days after iv. injections. INPs were found to accumulate more in CD31 staining structures 24 h after iv. INP injection when compared with 72 h and this difference was determined to be statistically significant. Over time, INPs were found to have accumulated inside more tumor cells and to cover more of the tumor cell surface. INPs injected iv. 3 days prior to RT appeared to increase median survival more than INPs injected 12 and 24 h prior to RT, although statistical significance was not reached. The fact that the INPs persist in tumors for at least 3 days post iv. injection without diminution suggests that INP-RT may be compatible with clinically relevant fractionated RT – that is currently being used and is under study [13–15]; INPs are shown herein to increase the efficacy of RT when provided as two 10 Gy fractions as well as one 20 Gy fraction. The study of the distribution of the INPs over time has relevance to other nanoparticle-based therapies and can be used to help determine the mechanisms by which these novel medications work.

Materials & methods

Iodine nanoparticles

The INPs were provided by (Nanoprobes, Inc., NY, USA) as an approximately 70 mg/ml solution. The physio-chemical features and detailed synthesis of the INPs have been described in a previous publication [12]. Briefly, the nanoparticles are polymerized by crosslinking the standard clinically used triiodobenzene contrast agent iohexol and covalently adding a 1K MW PEG coating. The particles have a hydrodynamic diameter of 19.6 nm with a polydispersity of 0.19 and a ζ-potential of -3.4 + 0.2 mV at pH 7.4. The blood half-life was found to be 40 h (2.7 days) and absence of toxicity at treatment doses determined by the absence of abnormal behavior, normal weight gain, the absence of abnormal clinical signs, normal complete blood work and normal histopathology of major organs. Since the iodine is covalently bound, no abnormalities in thyroxine or thyroid enzyme levels were found [12].

Tumor cells & cell culture

U87 human glioma cells, obtained from the American Type Culture Collection, were transduced with the pFULT vector, under the control of the human PGK1 promoter to express both luciferase and tomato, by the Skin Biology and Disease Resource-Based Center at Northwestern University (IL, USA), were grown in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum,1% L-glutamine (1.46 mg/ml), 1% penicillin–streptomycin and 1 mg/ml Fungizone.

Institutional animal assurance

Animal experiments were conducted according to NIH guidelines and approved by the institutional animal care and use committee of University of Connecticut Health Center and Stony Brook University before start of the study. UCHC protocol 101823-0521, RT, imaging and adjunct prophylaxis, was last approved 27 August 2018.

Mice & mouse housing

Six- to eight-week-old pathogen-free female athymic nude mice (~20 g) were purchased from Envigo (6903FHSD Athymic Nude FOXNLNU) and Charles River NCI (NCI-553 Athymic Nude NCR Nu/Nu) and were generally used within a week or two of receipt for these experiments. Mice were housed at the UCHC animal facility at a maximal density of five mice/cage.

Intracerebral tumor implantations

Mice were deeply anesthetized with intraperitoneal ketamine (140 mg/kg) and xylazine (3 mg/kg). After confirming deep anesthesia by toe pinch, a midline incision was made to expose the scalp. Using a hand-drill, a 0.5 mm wide burr hole was then drilled through the skull on the left coronal suture 2/3 of the way between the midline and the temporalis muscle insertion. Tumors were initiated by inoculating 1 μl of culture medium containing approximately 100,000 luciferase and tomato expressing U87 cells, 2.5–3.0 mm deep into the left hemisphere of the mouse brain through the burr hole using a 27-gauge injector connected to a 1-μl Hamilton microsyringe. Following a 1-min infusion of the cells, another 1 min was allowed for the cells to settle before removing the injector very slowly to minimally disturb the cells. The burr hole was then closed using surgical wax followed by skin closure using surgical glue.

Small & large tumors

Operationally, we defined small tumors to be approximately 1–2 × 10⁸ photons/s and large to be approximately 7–10 × 10⁸ photons/s.

Experimental grouping of mice

For therapy experiments, mice with approximately equivalently sized tumors based on IVIS measurements were sorted into experimental groups of approximately seven to ten mice each prior to RT – a control group that received no radiation, a RT group that received only RT and an INP + RT group that received INPs and RT under various conditions specified in the text and figure legends.

Intravenous Injections of the INPs

Mice receiving INP + RT received a total intravenous dose of either 3.5 or 7.0 g I/kg. INPs were iv. injected (tail vein) with a concentration of 70 mg I/ml INPs (two or four injections, two each day spaced >3 h apart). At 70 mg I/ml the INP solution was easily injected at a rate of about 0.5 ml over approximately 4 s using a 27-gauge needle without warming. Viscosity was acceptable at this concentration.

Mouse anesthesia

Mice were injected intraperitoneally with a ketamine/xylazine solution consisting of 0.0625 ml/20 g consisting of 1.1 ml phosphate-buffered saline (PBS), 1.0 ml Ketamine solution (100 mg/ml, Lederle Parenterals, Inc., Carolina, Puerto Rico) and 0.11 ml Xylazine solution (20 mg/ml, Ben Venue Labs, CT, USA).

Irradiations

Irradiations were administered to ketamine/xylazine anesthetized mice 12, 24 or 72 h after the last INP injection, as indicated in the figure legends. RT doses were either single dose 15 or 20 Gy or 2× 10 Gy given on consecutive days. Irradiations used a Philips RT100 x-ray generator (Amsterdam, The Netherlands) operating at 100 kVp with a 1.7 mm Al filter. Dose was calibrated using a Radcal ion chamber (CA, USA). The head and body were protected by a 3.4 mm-thick lead shield with a notch that enabled irradiation 8.0 mm caudally from the posterior canthus of the left eyelid and dorsally from the dome of the palate to above the calvarium. To prevent lethal brain edema, dexamethasone (5 mg/kg) was injected subcutaneously 18 h and again 6 h before irradiation and 6 h and again 18 h after irradiation [16]. Mice were euthanized when their weight fell below 80% of their weight on day of irradiation, or when they manifested signs of illness, for example – they began seizing, or were unable to immediately right themselves when they were flipped on their sides (flip test).

Cardiac perfusion/fixation

Twenty-four and seventy-two hours after iv. INP injections, mice were deeply anesthetized (unresponsive to toe pinch) and perfused trans-cardially – first with PBS and then with 10% buffered formalin (Fisher Healthcare, PA, USA). Formalin-perfused/fixed mouse brains were excised, incubated in 10% buffered formalin overnight at 4°C and then transferred to 30% sucrose solution for 24 h or until the brain sinks to the bottom of the solution. The brains were then cut coronally through the site of tumor implantation and arranged tumor-faced-down in cassettes containing Cryomatrix (Thermo Fisher Scientific, MA, USA) prior to rapid freezing in 2-methylbutane cooled in dry ice for 1 h. The cassettes were stored at -20°C prior to sectioning.

Immunostaining & immunofluorescence

Cryopreserved brain tissues were cut into 7 μm-thick coronal sections using a Cryostat (Leica, Cat. #: CM 3050S, Wetzlar, Germany) at approximately -24–26°C. 4,6-damidino-2-phenylindole 1:1000 was used to stain nuclei. Primary antibodies used were rabbit anti-PEG (1:500; Abcam Cat # AB512572, Wetzlar, Germany) and goat anti-CD31 (1:100; Abcam Cat # 19808). Secondary antibodies used were donkey anti-rabbit Alexa Fluor 488 (1:400; Invitrogen A21206, CA, USA) and donkey anti-goat Alexa Fluor 647 (1:200; Life Technologies A21447, CA, USA). Tissue sections were incubated overnight in primary antibodies. After washing four-times with PBS, secondary antibodies were added for 1 h. After washing four-times and coverslip mounting, imaging was performed using a widefield fluorescence microscope (Zeiss Axio Observer Z1) and a confocal fluorescence microscope using 10× and 63× objectives (Zeiss LSM 880).

In vivo imaging system

Mice were anesthetized during in vivo imaging system (IVIS) scanning using isoflurane aerosolized at a rate of 2 ml/min for periods of 20–25 min during scanning. An in vivo imaging system (IVIS Spectrum^®, Perkin Elmer, MA, USA) was used to quantify viable tumor cells. Imaging was done 13 min after a subcutaneous injection of luciferin (37.5 mg/kg) into isoflurane anesthetized mice bearing luciferase-expressing tumors. Photon counts were proportional to the number of luciferase transfected U87 tumor cells. A signal of approximately 8.0 × 10⁸ photon/s corresponded to a tumor of approximately 2.5–3 mm diameter, which was confirmed by inspecting the brains of U87-RedFluc tumor bearing mice.

MicroCT imaging & quantitation

Various times after iv. injection (INPs at 70 mgiodine/ml), mice were imaged by microCT (Scanco Medical VivaCT 80, Bruttisellen, Switzerland), operated at 70 kVp. The source spot size was 5 μm (with 0.5 mm Al filtering), sampling with 15 Å~ 15 ×15 μm voxels in a 30 mm-diameter field. 3 mm stacks of 200 sections at 2000 projections per revolution and an integration time of 300 ms/projection were collected, each stack requiring 20 min. Data were analyzed and viewed using Amira 3.1 software (Thermo Fisher Scientific). Standards were prepared in tubes filled with a range of sodium iodide and Iohexol concentrations. Quantification was done by averaging the intensity over tissue volumes and reading the value from the standard curve adjusted for uninjected tissue values.

Data analysis & statistics

Images were analyzed using Zen 2.3 Lite according to set criteria depicting different conditions looked for in the experiment. These included INPs inside the cell, INPs associated with CD-31 and INPs coating of the tumor cell. INPs were denoted as inside the cell when they were completely enveloped with the tumor signal. INPs were denoted as associated with CD-31 when there was a significant (++) or minor (+) colocalization with CD-31 signal. INP coating of the tumor cell was estimated as a percentage of the periphery of the tumor cell that was coated with INPs. The degree of CD31/INP overlap or INP coating of the tumor cell surface was determined by eye. Statistical significance was determined by using one- and two-tailed two proportion z-tests.

Results

Greater accumulation of iv. injected INPs in larger intracerebral U87 tumors compared with smaller tumors

U87 human glioma cells expressing both luciferase and tomato were implanted into the brains of athymic nude mice. Tumor progression was followed by IVIS. INPs were injected iv. when tumors were deemed small or large. At 24 and 72 h after INP injection, the mice were subjected to MicroCT and the resultant contrast in the tumors was quantified. Figure 1 shows MicroCT scans of small and large U87 tumors 24 h after INP injections (3.5 g/kg iodine) in two tail vein injections given ≥3 h apart. Greater contrast can be seen in the larger tumors. Iodine content was quantified and is shown in Table 1. Median iodine content was 0.17 ± 0.04 gI/kg small tumor versus 0.55 ± 0.04 gI/kg large tumor at 24 h (ratio 3.23) and 0.27 ± 0.08 gI/kg small tumor versus 0.6 ± 0.03 gI/kg large tumor (ratio 2.22) at 72 h.

Figure 1. . Larger intracerebral U87 tumors take up more iodine nanoparticles than smaller tumors as revealed by MicroCT analysis.

INPs (3.5 g/kg) injected iv. in a series of two injections separated by approximately 3 h. (A) Small tumor (IVIS: ∼1–2 × 10⁸ photons/s); (B) larger tumor (IVIS: 8–9 × 10⁸ photons/s). Shown are representative mouse scans from a pool of three each (Table 1).

INP: Iodine nanoparticle; iv.: Intravenous; IVIS: In vivo imaging system.

Table 1. . Quantification of iodine nanoparticles loading of smaller and larger tumors following iodine nanoparticle injections intravenous by MicroCT.

Tumor uptake 24 h after INP injections (% iodine by wt)		Small versus large tumor uptake ratio at 24 h	Tumor uptake 72 h after INP injections (% iodine by wt)		Small versus large tumor uptake ratio at 72 h
Small tumors	Large tumors		Small tumors	Large tumors
Mouse 1 = 0.14	Mouse 4 = 0.47		Mouse 1 = 0.26	Mouse 4 = 0.59
Mouse 2 = 0.16	Mouse 5 = 0.55		Mouse 2 = 0.23	Mouse 5 = 0.58
Mouse 3 = 0.21	Mouse 6 = 0.62		Mouse 3 = 0.31	Mouse 6 = 0.63
Average = 0.17	Average = 0.55	3.22	Average = 0.27	Average = 0.60	2.22

U87 human glioma cells expressing both luciferase and tomato were implanted into the brains of athymic nude mice. Tumor progression was followed by IVIS. INPs were injected iv. when tumors were deemed small (∼1–2 × 10⁸ photons/s) or large (∼7–8 × 10⁸ photons/s). At 24 h and 72 h after INP injection, the mice were subjected to MicroCT and the resultant contrast in the tumors was quantified as described in the ‘Materials & methods’ section.

INP: Iodine nanoparticle; iv.: Intravenous; IVIS: In vivo imaging system.

INP enhancement of RT may increase when INP accumulation is 72 h compared with 12 or 24 h

Intracerebral U87 tumors were allowed to reach a relatively large size (8–10 × 10⁸ photons/s) prior to RT. INPs (3.5 g I/kg, total dose) were injected as two tail vein injections given 3 h apart either 72, 24 or 12 h prior to RT. At the time of RT, tumors in all three groups had reached approximately the same size; each experimental group of approximately ten mice had approximately equivalently sized tumors. Figure 2 shows that mice that received INP injections 12, 24 or 72 h prior to RT all survived significantly longer than mice that received RT only. While there was no statistical difference between median survival of mice in the 12- and 72-h (p < 0.2) or 24 and 72-h groups, there was a trend toward greater survival in the 72-h group (median survival was 53, 57 and 65 days [1.3-fold, 1.46-fold, 1.79-fold] for 12, 24, 72 h, respectively).

Figure 2. . Iodine nanoparticle-enhanced radiation therapy was most effective when radiation therapy was administered 72 h after iodine nanoparticle injection when compared with 12 and 24 h.

Mice with intracerebral U87 tumors divided into five groups of equivalently sized tumors. The groups were control, no treatments, RT only (15 Gy, single fraction) and INPs (3.5 g/kg) administered 12, 24 and 72 h prior to irradiations (15 Gy, single fraction). Control, blue, n = 9; RT, red, n = 10; INPs 12 h, green, n = 9; INPs 24 h, orange, n = 10; INPs 72 h, yellow, n = 10. All of the treatment groups were significantly different from the control group; All of the INP treated groups were significantly different from the RT only group; none of the INP treated groups were significantly different from one another. INP treatment 72 h prior to RT trended better than 12 h prior to treatment, p < 0.2.

INP: Iodine nanoparticle; RT: Radiation therapy.

Fluorescence-based microlocalization of INPs in U87 tumors 24 & 72 h after iv. INP injections

Figure 3 shows a representative confocal image of the U87 tumor center 24 h after INP injection. INPs are seen to localize to vessel-like structures that contain bright CD31 stain. INPs also are found associated with the exterior of tumor cells with dimmer CD31 stain. Most of the CD31 stain (bright and dim) is co-localized with INPs. Images such as these were used to quantify INP and CD31 localization 24 and 72 h after INP injection. Figure 4 shows the confocal optical sections viewed from the side, in other words, visualizing the sections in the z direction. The upper panel (Figure 4A) depicts INP fluorescence on the tumor cell surface in close apposition to tumor cell nuclei. Distance of INPs from tumor cell nuclei was sorted into four categories: less than a half, one, one to two, or greater than two tumor cell nucleus diameters. The middle panel (Figure 4B) shows INPs that have been taken up by the tumor cell. The green INPs within the red fluorescing tumor-cells appears as a spherical yellow region (arrow), combining red from the tumor stain and green from the INP stain (individual red and green channels not shown). The lower panel (Figure 4C) depicts the coating of the tumor cells by the INPs. The amount of INP coating was quantified by counting tumor cells with either no INP coating, partial INP coating (0–10%, 10–25%, 25–50%, 50–100%). The results of this analysis for 24 and 72 h after iv. injection are shown in Figures 5–7.

Figure 3. . Fluorescently labeled coronal histological section of mouse brain bearing a U87 tumor.

(A) Iodine nanoparticles are labeled with anti-PEG (green), (B) endothelial stain anti-CD31 (violet), (C) tumor cells are expressing td-Tomato (red), (D) all colors together with nuclei (DAPI). White bar = 10 μm. Shown is a representative image from a pool of 63× images derived from three mice with advanced intracerebral U87 tumors.

Figure 4. . Criteria used for the semiquantitative analysis of iodine nanoparticle distributions on U87 tumor cells 24 and 72 h after intravenous injections.

INPs in proximity to the nucleus (A), INPs inside the cell (B), INPs associated with CD31 (not shown here), INPs coating the tumor cell surface (C). Confocal images (63×) were analyzed using Zen 2.3 Lite according to set criteria depicting different conditions looked for in the experiment. These included INPs inside the cell, INPs associated with CD-31 and INPs coating of the tumor cell. INPs were denoted as inside the cell when they were completely enveloped with the tumor signal. INPs were denoted as associated with CD-31 when there was a significant (++) or minor (+) association with CD-31 signal. The percentage of the tumor cell periphery coated by the INPs was estimated by inspection and then assigned to bins (Figure 6A & B).

INP: Iodine nanoparticle.

Figure 5. . Intracellular iodine nanoparticles (A) and iodine nanoparticles associated with CD31 (B).

(A) Quantitative analysis of the proportion of U87 glioma cells with INPs inside as a function of time after the iv. injection of INPs. (+), INPs inside the tumor cell; (-), INPs are absent from the tumor cell. n = 290; Images used were pooled from three mice; p = 0.008; z = ±3.34. INPs were denoted as inside the tumor cell when the yellow INP signal (individual red + green images not shown) was entirely surrounded by tomato signal. A cell with a single yellow spot was counted as (+). A two-tailed, two proportion Z-test was used. (B) Semiquantitative analysis of the association of INPs with the endothelial marker CD31. Shown are percentages of CD31 signal that has INP signal overlap 1 and 3 days post INP injection (iv.). Images used were pooled from three mice; p = 0.032; z = ±2.15. n = 141. A two-tailed, two proportion Z-test was used.

++: Complete overlap; +: Partial overlap; -: No overlap; INP: Iodine nanoparticle.

Figure 6. . Coating of tumor cell surface by iodine nanoparticles.

(A) Semiquantitative analysis of the percent of the tumor cell membrane coated by INPs. Arbitrary bins 0–25% (very little coating), 25–75% (moderate coating) and 75–100% (complete coating) were used to approximate the overlap of PEG staining (INPs) and the outer aspect of the tumor cell. n = 130; Images used were pooled from three mice; comparing 24 and 72 h: p = 0.109 (0–25%); p = 0.0526 (25–75%); p = 0.284 (75–100%)). An one-tailed, two proportion Z-test was used. (B) Semiquantitative analysis of the percent of the tumor cell membrane coated by INPs (A) using different binning. cells were reanalyzed to include new bins 0-<1%, 1–10%, 10–25%, 25–50%, 50–75% and 75–100%. n = 262; Images used were pooled from three mice; comparing 24 and 72 h: p = 0.099 (0-<1%), p = 0.309 (0–10%), p = 0.5 (10–25%), p = 0.036 (25–50%), p = 0.39358 (50–75%), p = 0.203 (75–100%). A one-tailed, two proportion Z-test was used.

INP: Iodine nanoparticle.

Figure 7. . Survival curves of mice with advanced U87 intracerebral gliomas treated with single fraction radiation therapy (1× 20 Gy) or fractionated radiation therapy (2× 10 Gy) with and without prior intravenous injections of iodine nanoparticles.

Groups of mice with equivalently sized tumors were injected with INPs when tumors reached approximately 8 × 10⁸ photons/s determined by in vivo imaging system. 1× 15 Gy: Mice received four injections of INPs over 2 days. RT was given 24 h after the last INP injection. 2× 10 Gy: Mice received 3.5 g/kg INP. RT (1× 10 Gy) was given 24 h later. 24 h later, mice received 3.5 g/kg INP followed by 1× 10 Gy 24 h after the final INP injection. In both 1× 20 Gy and 2× 10 Gy, RT post INP injections produced statistically significant prolongation of life compared with RT alone. (A) 2× 10 Gy, Survival; (B) 1× 20 Gy, Survival; (C) 2× 10 Gy, IVIS; (D) 2× 10 Gy + INP, IVIS; (E) 1× 20 Gy, IVIS; (F) 1× 20 Gy + INP, IVIS. Each colored line represents an individual mouse. A and B, no treatment control, n = 12; A, RT-only, n = 8; A, RT + INP, n = 8; B, RT-only, n = 7; B, RT + INP, n = 8.

INP: Iodine nanoparticle; RT: Radiation therapy.

(B), (E) and (F) reproduced with permission from [8], licensed with CC BY 4.0.

Semi-quantitative analyses of INP uptake & distribution on tumor cells 24 & 72 h after INP injections. Quantification of intracellular uptake of INPs into tumor cells 24 & 72 h after iv. INP injection

Figure 5A shows the quantification of tumor cells that have taken up INPs 24 and 72 h after iv. INP injections. The percent of tumor cells with intracellular INPs 24-h after INP injections was 38.2% (n = 290). After 72-h, 57.9% of tumor cells had intracellular INPs (n = 290). The increase in the percent of tumor cells with intracellular INPs was statistically significant (p < 0.0008).

Semiquantitative analysis of the association of INPs with the endothelial marker CD31

Shown in Figure 5B are the percentages of CD31 fluorescent signal that also has INP fluorescent signal associated/overlapping with it both 1 and 3 days after iv. INP injections. The degree of signal overlap was judged by eye: ++, complete overlap; +, partial overlap and -, no overlap. Most of the CD31 stain is completely colocalized with INP staining; approximately 73 and 87% of CD31 stain also had complete overlap with INP stain 24 and 72 h after INP injections, respectively. This increase is statistically significant (p < 0.03, n = 141).

Percent of the tumor cell membrane coated by INPs 24 & 72-h post iv. INP injections

Figure 6A is a graph demonstrating by increments the percent of the tumor cell membrane coated by INPs 24 and 72 h post iv. INP injections as determined by the overlap of PEG staining with the tumor cell membrane. The amount of the tumor cell membrane is given in increments of 0–25% (very little coating), 25–75% (moderate coating) and 75–100% (complete coating). One day post INP injection, about half of the tumor cell membrane had either no or very little INP coating; a third had moderate surface coating and approximately 10% had very extensive coating. By 72 h, nearly half of the tumor cells had moderate INP covering (p < 0.1, n = 65). Figure 6B dissected the binning further. In a separate analysis of 129 tumor cells each, at 24 h, approximately 6% of the tumor cells had no INP on their surface, 43% had low levels (1–10%), 30% had moderate levels (10–25%), 13% had 25–50%, 6% had 50–75% and 2% had 75–100% covering. At 72 h, 1.5% had no INP on their surface, 40% had 1–10%, 30% had 10–25%, 22% had 25–50%, 5% had 50–75% and 1.5% had 75–100% covering. With time, the number of tumor cells with no INPs and 0–10% INPs had decreased while those with 25–50% INP covering had increased.

Comparison of single-fraction 20 Gy RT for advanced intracerebral U87 gliomas & hypo-fractionated 2× 10 Gy RT: efficacy of INP-enhanced RT

Athymic nude mice with advanced U87 intracerebral gliomas were irradiated with and without prior injections of INPs (7 g/kg) given 24 h prior to the first irradiation. One group received two 10 Gy RT doses (2× 10 Gy) administered over 2 days (Figure 7A) while the other group received a single dose of 20 Gy RT (1× 20 Gy) (Figure 7B). Results are shown in Figure 7. INPs enhanced RT when mice received either 1× 20 Gy (p < 0.019) or 2× 10 Gy (p < 0.005). However, as expected a single 20 Gy dose provided greater life extension than 2× 10 Gy for both RT-only and RT + INP. For 1× 20 Gy, median survival was 40 days for RT only (twofold over control) and 170D for RT + INPs (4.25-fold over control). For 2X10Gy, median survival was 56D for RT-only (1.4-fold) and 79D for RT + INPs (twofold). Shown in Figure 7A–D are the corresponding IVIS measurements of tumor size as a function of time after irradiation. Decrementing tumor is seen in some of the mice receiving 1× 20 Gy.

Discussion

Intravenously injected INPs become concentrated into the U87 intracerebral glioma growing in the brains of athymic nude mice. In this paper we show that the amount of INPs that get into the tumors is related to the size of the tumor; larger tumors have 2–3× greater INP concentration than do smaller tumors (Figure 1 & Table 1). While we do not yet know the mechanistic basis for this observation, the result has a number of implications for INP-enhanced RT. Larger tumors are known to express more VEGF than smaller tumors, along with factors such as HIF-b and TGF which also trigger angiogenic pathways [17]. This results in a leakier BBB [18] and could result in increased INP accumulation. This is likely to be of clinical relevance. Larger INP accumulations and dose-enhancements might be expected in the main tumor masses while migrated tumor cells and smaller accumulations far from the main tumor mass might experience smaller INP accumulations and smaller dose enhancements. We have shown that gold nanoparticles selectively access tumor cells that have migrated far away from the main tumor mass using an invasive rat glioma model [7]. Further studies are needed to determine how INP redistributions over time affect INP access to migrated tumor cells in more aggressive, invasive glioma models growing in the brains of mice. Larger tumors may also undergo developmental alterations that promote INP accumulation. This is an area for further study. As an example, differently sized nanoparticles could be used to gain information about nanoparticle entry into tumors of different sizes during tumor development.

RT performed 72 h after INP injections showed a trend of being more effective than RT performed 12 and 24 h post INP injection (Figure 2). While the median 72-h versus 12 and 24-h survival did not meet statistical significance, the trend was clear (p < 0.2). If 7 g/kg INP had been used instead of 3.5 g/kg, statistical significance might have been reached as the INP-dose enhancement effect would have been greater [8]. Table 1 suggests more INP is loaded into the tumors after 72 h compared with 24 h. Waiting longer times after INP infusion before initiating RT might have significant benefits. Since the INPs circulate with a blood half-life of 40 h [12], blood levels would be lower and nonspecific endothelial irradiations would be minimized. MicroCT performed 72 h after INP infusions show essentially no decrement of INP levels. Further, INP redistributions over time, documented in this study, might benefit RT efficacy.

Some of the criteria used to analyze the distribution of the INPs on the confocal images are shown in Figure 4.

Intracellular INPs

Since the green INPs within the red fluorescing tumor cells appears as a spherical yellow region, tumor cells with clearly visible yellow regions were quantified. Figure 5 shows that by 72 h post INP injection, more tumor cells have intracellular INPs than after 24 h. The amount of INPs per cell was not quantified – but it is possible that tumor cells have more intracellular INPs/cell after 72 h as well. Since the ejected photoelectrons travel a short distance [6], it is possible that INPs that have accumulated into the tumor cells will result in more double strand breaks due to the possible proximity of the intracellular INPs to the cell nucleus or other sensitive cellular components such as mitochondria. It is possible that targeting the INPs to tumor cell receptors such as transferrin [19–21] will increase the intracellular accumulation of INPs due to receptor cross linking and endocytosis [22–24] thereby increasing intracellular uptake even more, with correspondingly greater efficacy.

Colocalization of INPs with CD31

Most of the INPs do not colocalize with CD31. However, there is almost complete overlap of CD31 staining with INPs. Colocalization increases with time after INP injections (Figure 5B). The full significance of the colocalization of CD31 with the INPs is not understood. CD31 is nominally an endothelial cell marker [25]. Indeed, most of the colocalized INPs can be found associated with vessel like structures that also contain bright CD31 stain. But INPs and less bright CD31 colocalized stains are also found associated with structures that do not look like classical vessels and may represent a form of vasculogenesis [26,27] or a novel form of vascular mimicry in which the CD31 and the INPs may be associated with either the tumor cells themselves or a structure very close to the tumor cell surface (Ridwan et al., unpublished observations).

Fraction of tumor cell coated by INPs

The amount of INP coating is quantified by counting tumor cells with either no INP coating or partial to total INP coating (0–10%, 10–25%, 25–50%, 50–100%). The fraction of the tumor cell surface coated by INPs increases with time (Figure 6A & B). By 72 h post INP injection, the percent of tumor cells with 50–75% of their surface coated by INPs increased from 15 to 30%. It is possible that a greater fraction of nuclei close to the tumor cell surface had INPs in direct apposition after 72 h but it is extremely challenging to quantify this since some aspect of the nucleus appears to be close to some aspect of the cell surface in almost every cell. Since ejected electrons are postulated to have greatest effect on nuclei close to cell membrane coated by INPs [6], waiting 72-h after INP injection should improve efficacy. Many of these hypotheses can be tested directly by quantifying the number of double strand breaks in nuclei [28] in relation to their degree of closeness to accumulated INPs.

Figure 7 shows that INP-enhanced RT is effective when RT is delivered as a single 20 Gy fraction (1× 20 Gy) or as two 10 Gy fractions (2× 10 Gy) given 1 day apart. Single fraction 20Gy RT is more effective than two 10 Gy fraction RT as would be expected; infusions of INPs prior to RT is sufficient to increase efficacy in both cases. This is important since standard RT is given clinically as fractionated RT given in 30 fractions of 2 Gy each. The potential benefits of delivering fewer numbers of larger dose fractions (hypofractionated RT) include better tumor control (greater cell kill and longer dormancy) as well as reducing treatment time. This is an active area of clinical research with many hypofractionation regimens in use [29]. It is likely that a single infusion of a clinically relevant dose of INPs prior to RT will be sufficient to provide enhanced RT efficacy for clinically-used hypofractionation schemes. The use of INPs to enhance fractionated RT needs to be studied in greater detail and in a variety of glioma models with different properties, as provided by well characterized PDX glioma models growing in the brains of immunocompromised mice [30] and syngeneic glioma models where an intact immune system and glioma growth/vascularization is obtained within a more physiologic context [31].

INPs bind at or near the tumor cell surface and are usually co-localized with the endothelial marker CD31 [8]. Further, INP-RT synergizes with DOXIL chemotherapy greatly increasing glioma survival [8]. Plausible mechanisms to explain these observations are suggested. It is possible that the INPs localized to tumor endothelium enhance endothelial damage, one of the barriers to drug delivery, especially in the brain. Studies confirm that endothelial/pericyte/interstitial extracellular matrix damage enhances vascular permeability and drug delivery [32–35]. The INPs may be particularly effective at increasing vascular permeability in the brain compared with RT only; this hypothesis needs to be tested [36]. Further, increased RT dose to the tumor vasculature may also decrease blood flow to the tumor more than that obtained by RT alone [37]. This hypothesis also needs to be tested. Further, the localization of the INPs to the tumor surface close to asymmetric nuclei lying close to the tumor membrane [6] should result in an increase in double strand DNA breaks in those nuclei [28]. Experiments quantifying double strand DNA breaks in and around advanced intracerebral U87 tumors as well as the normal brain are needed. Next investigative steps could therefore include experiments to identify complementary therapies that synergize with the INPs to greatly increase double strand DNA breaks in the tumor following RT and to exploit the ability of INP-enhanced RT to breakdown permeability barriers of brain tumors.

Conclusion

In conclusion, after iv. injection, INPs that are taken up into gliomas undergo a redistribution over time that might have significance for optimizing their powerful therapeutic efficacy as radioenhancers [8]. The excellent biocompatibility profile of the INPs [12] and the ability of INPs to access tumor cells far from the main tumor mass (Ridwan et al., unpublished observations) make the INPs a candidate for clinical translation to treat primary and metastatic brain tumors for which current therapies are largely inadequate.

Future perspective

INPs are very safe and effective radio-enhancing agents that not only extend median survival of mice with advanced intracerebral U87 gliomas when combined with RT but also synergize with chemotherapy making chemotherapy much more effective. Once the mechanisms by which the INP-RT are understood and INP tumor loading has been optimized and INP-RT has been shown to be safe and effective for the treatment of a larger animal such as dogs with spontaneous gliomas, human trials will be possible. Once a large therapeutic benefit to canine and human patients is demonstrated, demand will spur the development of state-of-the-art kilovoltage irradiators to take advantage of this technology that optimally enhances radiation therapy through the photoelectric effect.

Summary Points.

• Larger intracerebral U87 gliomas accumulate iodine nanoparticles (INPs) in higher concentration than do smaller tumors.

• INPs persist in U87 gliomas for at least 3 days following intravenous (iv.) injection and undergo changes in localization and distribution over this time.

• More CD31 staining structures were found to associate with INPs 72 h after iv. INP injection compared with 24 h.

• More INPs were found to accumulate inside tumor cells 72 h after iv. INP injection compared with 24 h.

• INPS were found to cover more of the tumor cell surface 72 h after iv. INP injection compared with 24 h.

• INPs injected iv. 3 days prior to radiation therapy (RT) showed a trend of increased median survival compared with INPs injected 12 and 24 h prior to RT.

• INPs are shown to increase the efficacy of RT provided as two 10-Gy fractions as well as one 20-Gy fraction.

• INP persistence and redistribution in tumors over time may enable greater RT enhancement and clinically relevant hypofractionated RT and may enhance INP efficacy.