USE OF RADIATION TO KILL CANCER BY NANOPARTICLES AND IN A BIODOSIMETER USING GENE EXPRESSION ANALYSIS

The use of radiation is broad in biological systems, in different areas of research mostly in health. Radiation is used to kill cancer. In radiation therapy proper calculation is done so that a maximum dose is delivered to the cancer . Inspite of this precaution, radiation effects healthy tissue. This effect is especially dangerous when the tumor is located near important organs. Thus in radiation therapy, it is important to reduce the dose and the damage to the healthy tissues and organs. The damage on the healthy tissues due to radiation therapy in cancer could be reduced by reducing the radiation dose to get the same treatment effect or by enhancing the radiation. The enhancement of radiation effect in vitro and in vivo can be obtained by targeted drug delivery on the cancer. Also photo dynamic therapy can be supplemented by the use of radiation therapy on the cancer by targeted drug delivery. Another use is the development of a bio dosimeter. In a large scale nuclear event it is important to measure the radiation dose exposed to humans. Also it is likely that the people who are exposed to radiation are not wearing the dosimeter. So a method of estimating radiation dose to a person exposed to radiation without a physical dosimeter would be a very useful procedure. One possible method is the use of gene expression analysis, which is based on the fact that the expression of the genes will change due to the absorbed radiation. So developing a biological dosimeter based on the gene expression analysis could quantify the radiation dose given to the patients during radiation therapy or to assess the risk of cancer developing in the general population. This biological dosimeter could even be used when the physical dosimeters are insufficient to estimate the risk caused by the radiation exposure or even years after being exposed to nuclear accidents. The main goal of the work presented here is to investigate the following topics Use of gold pHLIP to enhance the radiation effect in cancer cells Review on the in vivo research done to enhance radiation using gold nanoparticles Analyze the gene expression results from irradiated drosophila melanogaster to develop a biological dosimeter. Use of X-ray to activate targeted Copper Cysteamine nanoparticles photosensitizer to reduce tumor size in mice. A review work I have done on the researches related to enhancement of radiation using gold nanoparticles in tumor bearing mice showed that the targeted nanoparticles are a promising method for achieving radiation enhancement due to their shape, size , surface chemistry and the properties of the nanoparticles.. Gold nanoparticles are susceptible to X-rays compared to tissues and release extra electrons by Auger effect when the tumor treated with gold is irradiated. These auger electrons have low energy and are localized within the tumor site killing the cancer cells. However tumor targeting peptide pHLIP (pH Low Insertion Peptide) conjugated to gold nanoparticle specifically targets low pH medium (tumor) which when irradiated reduces the risk of killing healthy tissues near by and increases the uptake of the particles in the cancer mostly in the cellular membranes compared to only gold. The experimental results on cellular uptake of gold showed that there was an enhancement of gold uptake by 52% at low pH compared to normal pH ( P value = .008)and also in targeted gold by 34% compared to non targeted gold at low pH (P value= .023). The images obtained by distribution of gold experiment showed that the cellular uptake of gold-pHLIP is higher compared to gold alone . The targeting of plasma membrane by gold-pHLIP is seen clearly on all the images and some staining of internal organelles and nuclei membranes as well. The clonogenic assay experiment at 1.5Gray radiation showed a statistically significant 24% decrease in survival for cells treated with gold-pHLIP at low pH compared with cells treated with no gold. Also a statistically significant 21% decrease in survival for cells treated with gold-pHLIP at low pH compared with cells treated with gold alone. Thus Gold nanoparticles conjugated with pHLIP significantly increases the amount of gold particles in cancer cells thus enhancing the radiation effect and increasing the amount of cancer cell death from radiation. Copper cysteamine nanoparticles placed in the tumor site release cytotoxic singlet oxygen molecules on irradiation. The Cu-Cy nanoparticles being photosensitizers kill tumor when activated by radiation. Photosensitizers are limited to shallow tumors. Here we use X-radiation to photosensitize the pHLIP targeted Cu-Cy nanoparticles to kill even the deeply seated tumors in vivo. The results from the in vivo experiment we have done shows significant tumor destruction under X-ray activation. ANOVA analysis showed that the mice treated with targeted particles had a significantly different tumor sizes than mice treated with no particles, as well as mice treated with non-targeted particles. Also the use of targeted copper cysteamine nanoparticles affected the survival time after irradiation, compared to irradiation using no particles on mice. This work confirms the effectiveness of Copper Cysteamine nanoparticles, targeted to tumors, as a photosensitizer when activated by radiation therapy. Thus the aid of radiation therapy to photodynamic therapy by the use of tumor targeted CuCy nanoparticles efficiently does tumor destruction shrinkage with the increase in mice survival. Gene expression analysis on a published data showed that the expression of genes are radiation dose dependent and some genes behaving predictably as a function of radiation dose at different time points after radiation can be used as a bio dosimeter. The data analysis showed that 6 genes from drosophila melanogaster show linear response (R 2 > 0.9) with radiation dose at all time points after irradiation. Four of these genes have human homologues. Dropping off the lowest radiation dose (10 roentgen being very low for the fruit flies), 13 genes show a linear response with dose at all time points including 5 of 6 genes in whole data set. Of these 13 genes, 4 have human homologues and 8 have known functions. The Irbp (inverted repeat – binding protein) gene among the above is very important as it is a DNA repair gene. It is reasonable to predict that DNA damage is linear with radiation dose; thus, it is logical that some DNA repair genes may respond linearly in expression. Irbp has homologues in organisms that are as complex as humans and chimpanzees and in organisms as Japanese rice . The expression of this panel of gene, particularly those with human homologues, could potentially be used as the biological indicator of radiation exposure in dosimeter applications. Thus we could use radiation to kill tumors more effectively or the development of a biological dosimeter could help people to estimate the risk of cancer caused due to their exposure to radiation.

dosimeters are insufficient to estimate the risk caused by the radiation exposure or even years after being exposed to nuclear accidents.
The main goal of the work presented here is to investigate the following topics -Use of gold pHLIP to enhance the radiation effect in cancer cells -Review on the in vivo research done to enhance radiation using gold nanoparticles -Analyze the gene expression results from irradiated drosophila melanogaster to develop a biological dosimeter.
-Use of X-ray to activate targeted Copper Cysteamine nanoparticles photosensitizer to reduce tumor size in mice.
A review work I have done on the researches related to enhancement of radiation using gold nanoparticles in tumor bearing mice showed that the targeted nanoparticles are a promising method for achieving radiation enhancement due to their shape, size , surface chemistry and the properties of the nanoparticles.. Gold nanoparticles are susceptible to X-rays compared to tissues and release extra electrons by Auger effect when the tumor treated with gold is irradiated. These auger electrons have low energy and are localized within the tumor site killing the cancer cells. However tumor targeting peptide pHLIP (pH Low Insertion Peptide) conjugated to gold nanoparticle specifically targets low pH medium (tumor) which when irradiated reduces the risk of killing healthy tissues near by and increases the uptake of the particles in the cancer mostly in the cellular membranes compared to only gold.
The experimental results on cellular uptake of gold showed that there was an enhancement of gold uptake by 52% at low pH compared to normal pH ( P value = .008)and also in targeted gold by 34% compared to non targeted gold at low pH (P value= .023). The images obtained by distribution of gold experiment showed that the cellular uptake of gold-pHLIP is higher compared to gold alone . The targeting of plasma membrane by gold-pHLIP is seen clearly on all the images and some staining of internal organelles and nuclei membranes as well. The clonogenic assay experiment at 1.5Gray radiation showed a statistically significant 24% decrease in survival for cells treated with gold-pHLIP at low pH compared with cells treated with no gold. Also a statistically significant 21% decrease in survival for cells treated with gold-pHLIP at low pH compared with cells treated with gold alone. Thus Gold nanoparticles conjugated with pHLIP significantly increases the amount of gold particles in cancer cells thus enhancing the radiation effect and increasing the amount of cancer cell death from radiation.
Copper cysteamine nanoparticles placed in the tumor site release cytotoxic singlet oxygen molecules on irradiation. The Cu-Cy nanoparticles being photosensitizers kill tumor when activated by radiation. Photosensitizers are limited to shallow tumors.
Here we use X-radiation to photosensitize the pHLIP targeted Cu-Cy nanoparticles to kill even the deeply seated tumors in vivo. The results from the in vivo experiment we have done shows significant tumor destruction under X-ray activation. ANOVA analysis showed that the mice treated with targeted particles had a significantly different tumor sizes than mice treated with no particles, as well as mice treated with non-targeted particles. Also the use of targeted copper cysteamine nanoparticles affected the survival time after irradiation, compared to irradiation using no particles on mice. This work confirms the effectiveness of Copper Cysteamine nanoparticles, targeted to tumors, as a photosensitizer when activated by radiation therapy. Thus the aid of radiation therapy to photodynamic therapy by the use of tumor targeted CuCy nanoparticles efficiently does tumor destruction shrinkage with the increase in mice survival.
Gene expression analysis on a published data showed that the expression of genes are radiation dose dependent and some genes behaving predictably as a function of radiation dose at different time points after radiation can be used as a bio dosimeter.
The data analysis showed that 6 genes from drosophila melanogaster show linear response (R 2 > 0.9) with radiation dose at all time points after irradiation. Four of these genes have human homologues. Dropping off the lowest radiation dose (10 roentgen being very low for the fruit flies), 13 genes show a linear response with dose at all time points including 5 of 6 genes in whole data set. Of these 13 genes, 4 have human homologues and 8 have known functions. The Irbp (inverted repeatbinding protein) gene among the above is very important as it is a DNA repair gene. It is reasonable to predict that DNA damage is linear with radiation dose; thus, it is logical that some DNA repair genes may respond linearly in expression. Irbp has homologues in organisms that are as complex as humans and chimpanzees and in organisms as Japanese rice . The expression of this panel of gene, particularly those with human homologues, could potentially be used as the biological indicator of radiation exposure in dosimeter applications.
Thus we could use radiation to kill tumors more effectively or the development of a biological dosimeter could help people to estimate the risk of cancer caused due to their exposure to radiation.
vi ACKNOWLEDGMENTS I would like to express my deepest gratitude to my major adviser Dr Michael Antosh for providing me an opportunity to work in his laboratory as a first graduate student.

Introduction
Gold is an inert and generally non-toxic material with unique properties suitable for many applications such as cancer diagnosis and treatment (1-7). Nanometer-size gold particles have recently been shown to increase radiation damage to tumors (2,(8)(9)(10)(11).
With enhanced radiation, the same level of tumor killing can be had with less radiation to a patient, reducing side effects of radiation treatments. Similarly, more tumor killing can be had for the levels of radiation that are currently given.
The increase in radiation effectiveness with gold nanoparticles is due largely to two causes. First, gold is capable of absorbing radiation at a significantly higher rate than tissue, up to about 100 times more for keV energies (2 (12)(13)(14). In addition, histological studies showed that antibodies conjugated with gold nanoparticles do not penetrate deeply into tumors, but mostly stain peripheral tumor regions (15). The direct injection of micron-sized gold particles does not lead to tumor targeting, as particles stayed only at the injection site and were not able to diffuse even within a tumor, hindering tumor coverage (16).
Our approach is based on targeting of tumor acidity, which correlates with tumor malignancy (17)(18)(19). The pH-sensitive targeting agents we are developing are based on the action of a family of pHLIPs (pH Low Insertion Peptides), which can "sense" acidity at the surface of cancer cells and deliver diagnostic and therapeutic molecules to tumors of different origins (20)(21)(22)(23)(24)(25). It was shown that pHLIP can promote fusion of liposomes with cancer cells and cellular delivery of various payloads (26,27) including small gold nanoparticles (26). Recently, pHLIP was successfully employed for the targeting of various nanoparticles to tumors and other acidic diseased tissue (28)(29)(30)(31).
pHLIP has also been used to mediate pH-controlled delivery of both 13 nm water soluble gold nanoparticles coated with luminescent europium into human platelets in vitro (32), and 1.4 nm gold nanoparticles to tumors (33). Intratumoral and i.v administrations of both demonstrated a significant enhancement of tumor uptake of 1.4 nm gold nanoparticles conjugated with pHLIP. Statistically significant reduction of gold accumulation was observed in acidic tumors and kidney when pH-nonsensitive K-pHLIP was used as a vehicle, suggesting an important role of pH in the pHLIPmediated targeting of gold nanoparticles.
In this work, we made another important step toward clinical application of 1.4 nm gold nanoparticles conjugated with pHLIP. We show that pHLIP can deliver gold to cellular components in a pH-dependent manner and enhance the radiation damage in cells.

Results
In this work we used 1.4 nm diameter gold clusters functionalized with maleimide.
Maleimide-gold clusters were conjugated with WT-pHLIP containing a single Cys residue at the N-terminus: ACEQNPIYWARYADWLFTTPLLLLDLALLVDADET After conjugation, the construct was purified, lyophilized, redissolved in DMSO, quantified and used in experiments with cells. As a control (gold alone) we used nonfunctionalized 1.4 nm gold clusters.

Cellular Uptake and Distribution of Gold
We investigated uptake of gold nanoparticles at normal and low pHs (pH 7.4 and 6.0, respectively), with and without pHLIP on human lung carcinoma (A549 cells). At pH 6.0 pHLIP was found to increase cellular uptake of gold nanoparticles by 34% compared to gold nanoparticles alone (p value 0.023) ( Fig.1 and SI Appendix, Table   S1). The uptake of pHLIP-gold at pH 6.0 increased by 53% compared to the uptake at pH 7.4 (p value 0.008). The uptake of gold alone was also enhanced at pH 6.0 compared with pH 7.4 (P value = 0.014). The uptake of gold-pHLIP was ~60% of the treated dose (1.8µg), which was about 1.1µg gold. Because each treatment had ~1 million cells, the amount of gold per cell was ~ 1.1X10 -6 µg. We expect that uptake of pHLIP-gold at normal pH by noncancerous cells will be much lower, since pH at the surface of glycolytic cancer cells is about 6.6-6.8 even when bulk pH of media is 7.4 (unpublished data). The uptake of gold alone was also enhanced at pH 6.0 compared to pH 7.4 (p value 0.014).
Light microscopy was used to establish the distribution of gold nanoparticles in cells.
Bright field images of cells treated with gold-pHLIP or gold alone and enhanced with silver are shown in Figure 2. The cellular uptake of gold-pHLIP is higher compared to the uptake of gold alone (Figure 2A and -B; the images are taken using 20x objective).
The representative bright field image of cell treated with gold-pHLIP and enhanced with silver obtained at high magnification is shown in figure 2C (the image is taken using 100x objective). The overlay of fluorescent images of nuclear stained with DAPI (blue) and cellular membrane stained with HQ silver deposited on the goldnanoparticles (red) are shown in figure 2D. The targeting of the plasma membrane by gold-pHLIP is clearly seen on all images. We also observed some staining of internal organels and nuclei membranes. Targeting of mitochondria and nuclear membranes was observed in experiments with pHLIP-coated liposomes containing lipids conjugated with fluorescent dyes and gold nanoparticles (27).

Clonogenic Assay
Clonogenic assay experiments were performed to assess cell survival after treatment of cells with gold or gold-pHLIP and radiation of treated and non-treated cells. The results of the experiments are summarized in figure 3 and SI Appendix, Tables S2-S5. We tested 0, 1.5 and 3 Gray of radiation. Gold nanopoarticles alone or conjugated with pHLIP were not toxic for cells in the absence of radiation. For 1.5 Gray of radiation, we observed a statistically significant 24% decrease in survival for cells treated with gold-pHLIP at low pH, as compared to cells treated with no gold. We also observed a statistically significant 21% decrease in survival for cells treated with gold-pHLIP at low pH as compared to cells treated with gold alone. The effect of gold was not significant at 3 Gray of radiation, likely because the survival of cells at 3 Gray was low.
Two different methodologies were used: excess of gold or gold-pHLIP was removed after treatment with cells before radiation, or excess of gold and gold-pHLIP was not removed (non-removal corresponds with the values shown in red in SI Appendix ,   Tables S2-S5). The clonogenic assay results in Fig. 3A include data obtained at both different methodologies. Fig. 3B shows the data obtained in the experiments when gold constructs were not removed before radiation. Surprisingly, overall the nonremoval data have better survival than the removal data; perhaps this is a result of the removal process stressing the cells.
We assessed statistical significance for data obtained at 1.5 Gray of radiation by performing an analysis of variance (ANOVA), summarized in Table 1 and SI Appendix, Table S6. When determining the p values between different gold treatments, we accounted for the difference in methodology as an additional variable in the analysis of variance (see Methods section for more details). Our data clearly indicate that cell treatment with gold-pHLIP results in a statistically significant decrease in cell survival as compared to a treatment with no gold (p value 3.610 -5 ) or gold alone (p value = 0.015).
In a separate experiment, cells were treated with gold constructs at pH 7.4, where pHLIP is less effective at inserting into the cellular membranes. Only small and statistically insignificant differences in survival between non-treated and treated cells were seen; the data is given in SI Appendix , Tables S2, S3, and S5.

Discussion
The treatment of cancer involves a trade-off between killing all cancer cells and impacting healthy tissue and organs as little as possible. To reduce side effects and enhance lethal effects of radiation for cancer cells the approaches of binary therapy were introduced. Binary radiation therapy targets cells at the biological level with a noncytotoxic agent that is "activated" by low energy radiation, thereby destroying cancer cells wherever they may reside, while sparing normal cells in proximity to the diseased cells. A number of binary radiation therapies have been and are being explored (9,(34)(35)(36) one of the more promising approaches is based on dose enhancement through Auger electron emission secondary to the photoelectric effect dominant at low photon energies. Auger electron emission generates a cascade of lowenergy electrons that travel very short distances and deposit their energy locally. The number of Auger electrons generated in targeted cells can be increased significantly by introducing material of a high atomic number (high-Z) into the target as long as the radiation energy is at or near the K, L, or M electron shell binding energies for the material. High-Z nanoparticles made of iodine, gadolinium, or gold are predicted to produce a clinically achievable dose enhancement of as much as 10 fold. Because low energy electrons travel very short distances, it is crucial to deliver and accumulate high-Z material on or in cancer cells in tumors.
Our strategy is to deliver gold, which is an inert, high-Z material widely used in medicine, to cancer cells for enhancement of radiation effects. The delivery approach we propose is based on the energy of membrane-associated folding of peptides from the pHLIP family to target cellular membranes in a pH-dependent manner (22,24,37).
At pH <7.0 pHLIPs insert into the lipid bilayer of the membrane, which is accompanied by a coil-helix transition and formation of a transmembrane helix. It has been shown that pHLIP delivery agents can target acidic tumors with high accuracy and deliver nanoparticles, including gold, to cancer cells in tumors (33). In this work, we show the effect of gold-pHLIP on radiation-induced cell death.
The dose enhancement depends strongly on the photon energies used for irradiation, as well as on the location and the size of gold nanoparticles. Regarding the photon energy, the ratio of gold absorption to human absorption is highest between ~10 and 100keV, with the ratio reaching approximately as high as 100 (2). We used 250 kVp X-rays with Sn-Thoraeus filtering to use the high relative absorption by gold while also accounting for the fact that lower-energy photons will be absorbed at too small of a depth to be useful. Regarding the location, it is very important to deliver gold nanoparticles as close as possible to cancer cells, as the dose deposited by Auger electrons increases as distance from the gold nanoparticles decreases (11). We used pHLIP to locate the gold nanoparticles to cancer cells. Regarding the nanoparticle size, it is best to use as small a gold nanoparticle size as possible to minimize the energy deposited inside the gold by Auger electrons. Simulations by McMahon et al. (11) predict an increase in relative biological effectiveness for decreasing sizes of gold nanoparticles. We used 1.4-nm -diameter gold nanoparticles.
The results of our present study indicate that pHLIP causes cells to take up more 1.4nm gold nanoparticles than cells without pHLIP. The gold nanoparticles deposited by pHLIP mostly accumulate on the plasma membrane. As a result, gold nanoparticles delivered to cells by pHLIP can enhance radiation-induced decreases in cell survival.
Gold nanoparticles tethered to the lipid bilayer of the plasma membrane by pHLIP may trigger cell death by inducing oxidation of lipids, cholesterol and membrane proteins. The oxidized lipids are known to modify membrane physical properties, such as thickness, permeability, level of hydration and polarity, lipid transbilayer diffusion, loss of lipid asymmetry and phase segregation, which results in apoptosis (38,39).
The exposure of phosphatidylserine lipids to the outer leaflet of the lipid bilayer, promoted by lipid oxidation, serves as a recognition signal for macrophages to phagocytose the apoptotic cell (40).
The combination of the clonogenic and uptake results suggests that pHLIP is able to enhance radiation-induced death by targeting cancer cells and increasing gold uptake.
Monomaleimido nanogold was conjugated to Cys-pHLIP in 40 mM phosphate buffer containing 300 mM NaCl at pH 6.5. A reducing agent, Tris-(2-carboxyethyl) phosphine, was added into the reaction mixture (10x excess compared to pHLIP) to reduce pHLIP-S-S-pHLIP dimers and promote reaction with gold-malemide . The reaction vial was incubated overnight at room temperature on shaker. The next day, the gold-pHLIP conjugates were purified using Amicon Ultra (10K) centrifugal filters according to company recommended protocol. The product was then lyophilized, After 1 hour of treatment, the cells were pelleted using centrifugation (2,000 rpm x g for 5 min) followed by removal of treatment and washing cells with PBS three times.
The cells were then dissolved in concentrated nitric acid followed by sonication for about 2 hours. Concentrated solution samples were diluted to give 2%(wt/vol) nitric acid and analyzed via inductively coupled plasma mass spectroscopy (Thermo-Scientific x7 series) against calibration standards (IMS 103; UltraScientific). with 0.3% Triton X100 for 5 min, followed by washing with PBS and deionized water. Next, the cells were developed with freshly prepared HQ Silver reagent (Nanoprobes) for about 20 min, followed by washing with deionized water. Finally the cells were stained with 5 µM DAPI in PBS for 5 min, followed by washing with deionized water. The cells were imaged using light microscope in bright field regime to visualize gold enhanced by silver, and in the fluorescent regime to monitor DAPI and silver fluorescence using cut off filters (ex:em 360 nm/460 nm andex:em 542 nm/620 nm, respectively).  [12][13][14][15][16][17] and then reseeded in a 6-well plate.

Irradiation of Cells
Two hundred cells per well were seeded for 0-and 1.5-Gray radiation doses and 500 cells per well were seeded for 3-Gray radiation dose. In generally, six wells were seeded per treatment type; the number of entries in SI Appendix, A colony was defined as a distinct group of cells that contained 50 or more cells.

Analysis of Cellular Uptake Data:
The values in SI Appendix, Table S1 are six readings from a mass spectrometer. P values for statistical significance were computed using the t test, because the between-reading variance was much greater than the error in each reading.

Analysis of Clonogenic Assay Data:
To calculate statistical significance, the data summarized in SI Appendix, Tables S2-S5 were analyzed using an ANOVA, followed by a post hoc test using Tukey's Honest Significant Difference. Each individual measurement from the clonogenic assay dish was treated as a biological replicate, and normalized to the average of the "0 Radiation, No Gold" measurements from the same experiment.
The linear model fitted for the ANOVA had three variables: normalized survival (dependent variable), gold treatment (independent variable) and removal/non-removal of excess gold (independent variable). We left the data for 0 Radiation out of the analysis because normalizing by the "0 Radiation, No Gold" data points introduces a correlation if we use the data by which we are normalizing. We analyzed the data for 1.5 Gray and 3 Gray separately because we were only really interested in the effect at 1.5 Gray.
In the 1.5 Gray data, the interaction term between gold treatment and removal/nonremoval was significant. This is consistent with the gold treatments having different effectiveness depending on whether or not the excess gold was removed.
We also did one experiment at high pH, as mentioned in the results section. This was analyzed with a separate analysis of variance. The data are included as experiment 11 in SI Appendix, Tables S2. S3, and S5.
In Fig. 3   Error shown is SEM. Table S1. Cellular uptake of gold-pHLIP and gold. The mean and standard deviation of these data points are presented in figure 1, and the calculation of statistical significance is described in the methods section. Data are normalized so that Gold Alone, pH 7.4 has a mean of 1.

Introduction
In radiation therapy for cancer, radiation is delivered after precise calculations so that a maximum dose is given to the tumor and a minimum dose is given to healthy tissue.
Despite these efforts, radiation still affects healthy tissue. This effect is especially dangerous when the tumor is located near important organs. Thus, it is important in radiation therapy to reduce the dose and the damage to healthy tissues and organs(1).
One of the current strategies to reduce radiation is the use of radiation enhancers, which can absorb and make tumor cells more susceptible to it. They are designed to improve tumor cell killing, since making a tumor more susceptible to radiation means that less radiation can be used. And if less radiation is used, there will be less adverse effects on normal tissues (2).
Radiation enhancers can include materials like nanoparticles, e.g. carbon nanotubes, gold nanoparticles and quantum dots. In this paper, we will focus on the use of gold nanoparticles (GNPs). Gold is a good radiation enhancer. The radiosensitization of biomolecules by GNPs can be caused by locally increased radiation absorbed energy.
Gold, a high Z material, is capable of absorbing radiation at significantly higher rates than tissue. The advantage in absorption can grow to about a factor of 100 for certain keV photon energies (20 keV shown in (3), can be checked in a database (4)).
Additionally, gold nanoparticles that interact with radiation can release a number of short range also indicates that gold nanoparticles need to be located within tumors, near the vital cellular structures, in order to maximize the radiation enhancement effect (7,8). This suggests the need for the gold nanoparticles to be targeted to cancer cells.
In this paper, we will review the use of gold nanoparticles as a radiation enhancer in vivo. Specific topics include:  Properties of gold nanoparticles

 Important Experimental Variables
 In vivo radiation enhancement results for non-targeted gold nanoparticles  Nanoparticle Targeting  In vivo radiation enhancement results for targeted gold nanoparticles

Properties of Gold Nanoparticles
Gold nanoparticles properties include the following: i) Gold is an inert material and can be made to be biocompatible using surface modification, like surface coating of the GNPs (2, 3, 6-17).
ii) Gold nanoparticles can be linked to biomolecules, either via stabilizers (polyethylene glycol, maleimide) or directly to sulfhydryl (-SH) groups of moieties such as peptides, antibodies, small molecules or proteins.
iii) Gold nanoparticles, as any nanoparticle, have a large surface to volume ratio. The relatively large surface area provides opportunity for interactions with molecules. Having large number of surface ligands, gold nanoparticles allow flexible design and multi-functionality by incorporating mixed ligands (3,12,18) .
iv) The nanoparticles including GNPs exhibit preferential deposition at tumor sites due to the enhanced permeation and retention (EPR) effect.
This makes them to be effective as drug carriers and radiation enhancers (1,8,16). This is related to the small size of GNPs, and the leaky vasculature of tumors.

Important Experimental Variables
The following variables are known to affect the amount of radiation enhancement that the gold nanoparticles are capable of delivering: nm GNPs were widely distributed in different organs of the body due to small structures. Histological analysis showed that GNPs had almost no effect on tissues including liver, spleen, kidney, lung and heart, indicating good tissue biocompatibility of the GNPs.
Cell Line Used in Studies with GNPs: Radiation enhancement by GNPs is cell line specific. They enhance the radiation when treated with some cells but not all.
Significant radiosensitization occurred in MDA-MB-231 cells at 160 kVp. However, no significant radiosensitization was observed in DU 145 or L132 cells, even though there was uptake of GNPs in both of these cell lines. In an in vitro experiment, uptake of GNPs was greater in MDA-MB-231 cells than in DU 145 or L132 cells, and hence radiation enhancement was better in MDA-MB-231 cells. (6).

Intracellular Localization of GNPs:
The location of gold nanoparticles inside of the cells affects radiation enhancement; for example, a GNP attached to DNA (Deoxyribonucleic acid) will likely have a greater impact than a GNP in other locations (for example, the local effect model discussed in (25)). Typically, not targeted nanoparticles will enter cell via endocytotic pathway and will be trapped in endosomal/lysosomal compartments and might exit cell via the exocytosis process.
The uptake and removal of particles depend on its size, shape and surface properties (26). The use of pH Low Insertion Peptides (pHLIP ® peptides) to target gold nanoparticles to cancer cells (in vitro) resulted in location of GNPs to the plasma and nuclear membranes (7,15) . Ultra-small Au@tiopronin nanoparticles (2 and 6 nm) were localized throughout the cytoplasm and nucleus of cancer cells in vitro and in vivo, whereas 15 nm nanoparticles were found only in the cytoplasm and were aggregated (24).
The targeting ligands: The targeting ligands enable nanoparticles to bind to cell surface receptors and enter the cells by receptor mediated endocytosis (18).
Nanoparticles accumulate at the tumor sites due to leaky, immature vasculature due to enhanced permeability and retention effect (8,13,16). Chattopadhyay et al. (27) discusses the molecular targeting approach, which enables a larger amount of GNPs to cross the cellular membrane and accumulate in the cancer cell cytoplasm. The experimental result showed that the GNPs modified with trastuzumab for targeting HER-2 on breast cancer cells with 100kVp x-rays were more effective in decreasing the clonogenic cells survival as compared to the non-targeted GNPs (27). Kong et al. (28) found that the local concentration of GNPs in target locations can be increased by as PEG or BSA helps GNPs to avoid reticuloendothelial system uptake and to increase circulation time in blood (3,8). BSA capped GNPs are easy to synthesize, resulting in uniform size and stability under physiological conditions (8). A nonexhaustive list of similar or related methods includes the following:  Kim et al. (29) found that PEG-coated GNPs had a much longer blood circulation time (>4 h) than non-PEG-coated GNPs.
 PEG coated GNPs can accumulate in mouse sarcoma flank tumors to concentration 10 times that of muscle and 50 times that of brain (12).

Method of Administration of GNPs to Animals: Direct injection of GNPs by
intra-tumoral administration can aid tumor uptake (15). Intravenous administration of gold nanoparticles still results in relatively large accumulation in tumor tissue, due to the enhanced permeability and retention effect discussed above, which is related to leaky vasculature within tumors (see for example (14)).
The relative success of lower energy photons is likely due in part to the fact that, in general, lower energy photons have a higher absorption probability in gold than higher energy photons (4). However, lower energy photons also come with the complicating factor that they are less penetrating, and may not be able to reach tumors deeper than skin depth. For most clinical purpose MeV photons are used due to the fact that for high energy photons , the energy is distributed over a wide range in soft tissue (9). Liu et al. (36) found that the survival fraction for HeLa cells when irradiated with high LET carbon ions was significantly less than when irradiated with low LET X-rays.

In vivo radiation enhancement results for non-targeted gold nanoparticles
Although many gold nanoparticle radiation enhancement studies have been done in vitro, only a few studies have been performed in vivo. No specific tumor targeting was utilized in the studies described in this section. Other results showed that after injection of GNPs, many blood vessels became visible due to the gold absorption. Pharmacokinetics showed an early rapid rise followed by a slower clearance rate. Gold in tumor peaked at 7.0 ± 1.6 min and fell to half of its peak Similarly, for a slightly higher radiation dose of 35 Gy, 56% long-term survival was found compared to 0% survival of mice with no treatment and 18% long-term survival for radiation only.
Further results showed that IV injected GNPs specifically localized in brain glioma in a 19:1 tumor to normal brain ratio. The micro CT measured by the tumor uptake of 1.5 ± 0.2% (weight by weight) gold, which was considered to be the highest gold The results showed a retarded tumor growth and increase in survival of mice receiving GNPs followed by radiation compared to the radiation alone, GNPs alone and control groups of mice. Survival of mice treated with gold nanoparticles and radiation was 60% after 2 months, whereas survival was less than 20% for radiation treatment alone and 0% for gold nanoparticles alone or no treatment.
Biodistribution of GNPs 24 hours post IV injection of GNPs showed the accumulation of GNPs inside the tumor, with a tumor to tumor surrounding muscle gold ratio of 6.4:1. Also, higher concentrations of GNPs were found in the liver and spleen, indicating uptake of gold by the reticuloendothelial system.
The number of apoptotic cells detected in tumor by a TUNEL assay was significantly higher in mice treated with GNPs followed by radiation than in mice receiving only radiation, GNPs alone and control groups. suggesting that radiation-induced blood brain barrier disruption can be leveraged to improve the tumor-tissue targeting of GNPs, which would further optimize the radiation enhancement of brain tumors by GNPs. The GNP toxicity in vivo was very small in this study, as shown by the preliminary data. GNPs. This analysis showed no toxicity.

Nanoparticle Targeting:
Successful targeting increases the likelihood that each gold nanoparticle will reach the tumor. Thus, there is the potential for targeted gold nanoparticles to improve the radiation enhancement effect. This is particularly true when the primary benefit from gold nanoparticles comes from Auger electrons, which have a short range, as discussed in the introduction. In addition to locating gold nanoparticles to cells, the resulting intercellular localization is also important, as discussed above. Targeting strategies can be divided into two categories: those that use cancer-targeting molecules, and other methods that do not.

Cancer cell targeting molecules:
More specific tumor targeting can be done by surface conjugation (attachment) of antibodies, peptides and other tumor targeting molecules (12). This can improve the therapeutic index (16,40). Conjugating gold nanoparticles with targeting molecules enhances the interaction of the GNPs with the cell surface by enabling the GNPs to bind to the cell surface receptors and enter cells by receptor-mediated endocytosis (18). A non-exhaustive list of targeting molecules used with gold nanoparticles (either in vitro or in vivo) includes the following strategies:  Glucose capped GNPs are designed to take advantage of an increased cancer cell requirement for glucose in order to target the cell cytoplasm (28).  Shah et al. (46) found that 30 nm PEG-coated GNPs interact with blood cells in vivo, which results in longer blood circulation that correlates strongly with tumor uptake. In tumors, accumulation was increased by 10 times using GNPs conjugated with a bioactive ligand (tumor necrosis factor) compared to untargeted GNPs.

In vivo radiation enhancement results for targeted gold nanoparticles
To the author's knowledge, there are only a few papers currently existing where targeted gold nanoparticles are used to enhance radiation effects on tumor in vivo.  radiosensitized GNPs was done. Decreased vasculatization in tumors was seen after 1 and 6 weeks of the treatment in CTX-GNP+RT group than control, RT only and CTX+RT groups. Also Tunnel assay results showed that apoptosis was higher after 1 week and less apoptosis after 6 weeks of treatment in CTX-GNP+RT group compared to RT only group. And other results showed that the level of proliferation and tissue repair was reduced in CTX-GNP+RT group compared to other groups.. Further, no cytotoxic effect was seen on the mice.

Conclusion
The papers reviewed in this article demonstrate the potential effectiveness of gold nanoparticles in the enhancement of radiation of tumors. Major results and methodologies are summarized in Table 1. Future experiments with gold nanoparticles and radiotherapy will likely involve the following areas: trials in humans, experiments using targeted gold nanoparticles and different radiation energies/types.
The eventual goal of gold nanoparticle treatments is to become viable for use in humans. One potential roadblock is that treatment with kilo-voltage x-rays is only capable of penetrating human tissue to a shallow depth. Perhaps trials using this treatment could be done starting with melanoma or other tumors, which could be accessed via catheterization, and future advances in engineering could help to eliminate this roadblock.
The roadblock mentioned in the paragraph above may also inspire more work with different radiation energies and radiation types. For example, the result of Kim et al. (35) with protons seems particularly promising. Additionally, in vitro studies (6,10) have shown radiation enhancement with gold nanoparticles and higher energy photons, although the enhancement is generally somewhat less than kilo-voltage photon results.
Regardless of the radiation type, it appears that tumor targeting will be of great use in this type of therapy. To conclude, we can say that GNPs modified with tumor targeting agents as pHLIP, cetuximab, cRGD and trastuzumab successfully enhanced the radio sensitization of GNPs which can lead to more effective clinical radiotherapy with less toxicity in near future (7,16,27,33). Additional trials with other targeting methods would be beneficial and important.
In summary, gold nanoparticles are a promising research area with the potential to reduce the amount of radiation necessary in cancer treatments. Successful experimental work has already been done in this area, including work in mammals.
More work is needed, and this future work has the potential of pushing the field into clinical relevance.  genes showed a linear response (R 2 > 0.9) with dose at all time points. One of these genes, Irbp, is a known DNA repair gene and has a human homologue (XRCC6). The lowest dose, 10 R, is very low for fruit flies. If the lowest dose is excluded, 13 genes showed a linear response with dose at all time points. This includes 5 of 6 genes that were linear with all radiation doses included. Of these 13 genes, 4 have human homologues and 8 have known functions. The expression of this panel of genes, particularly those with human homologues, could potentially be used as the biological indicator of radiation exposure in dosimetry applications.

Introduction
In the occurrence of a large-scale nuclear event, such as those at Hiroshima, Nagasaki, Chernobyl and Fukushima Daiichi, the measurement of radiation dose in exposed humans can be of crucial importance to survival (1, 2). However, in this situation it is very likely that many people who are exposed will not be wearing dosimeters. Thus, a method of estimating radiation dose to a patient without a dosimeter would be a very useful procedure.
One possible methodology for this procedure is the use of gene expression [polymerase chain reaction (PCR), gene sequencing, microarray analysis, and other methods]. The hypothesis is that the expression of genes will change due to the absorbed radiation, and that this change can aid or even substitute for physical dosimeters and act as a biomarker to estimate the distributed dose or the overall exposure. It also helps then to predict the long-term risks of both acute and chronic exposure (3)(4)(5)(6).
In addition to not requiring equipment, such as a dosimeter, another potential advantage of a gene-expression dosimeter is the time scale over which the measurements can be made. Even after the radiation exposure has taken place, the biological indicators for bio dosimetry can still be determined. This would certainly be an advantage compared to the physical dosimetry (7). Some biodosimetric techniques could be used long times after exposure (from 6mths to more than 50 years) which makes it unique compared to the requirements for methods used for immediate dose estimation (8).
Biological dosimetry not only provides information about the range of radiation dose but also along with this provides information about the individual radio sensitivity, which depends on age, smoking habits or other environmental toxins. Thus, biological indicators are also a measure of the biological, medical radiation damage. Hence, we can predict about the possible radiation damage by the determination of biological indicators (5,7,9,10).
The possibility of using gene expression changes has been an exciting method to measure and predict the damage due to ionizing radiation. The exposure of cells or animals to ionizing radiation may cause DNA damage and trigger the highly complex molecular response, resulting in changes of gene expression. These molecular responses may provide the prospective indicator of exposure (1,3). Previous work in this area showed that the variation in the response of genes is due to dose, dose rate, radiation quality and time after radiation exposure. This suggests that gene expression analysis may be an informative marker of radiation exposure and hence can be used as a potential biomarker. It is important to understand the cellular response to ionizing radiation or biological effects of radiation exposure in order to develop the predictive markers for the risk assessment due to radiation exposure on humans(1). The rigorous research going on in genomics and bioinformatics enables the development of gene expression profiling as a useful biological indicator of radiation exposure (10,11).
Work on this area until now has shown that the fold change in gene expression in response to radiation must be measured directly to develop a gene expression biomonitor. The expression of the genes would then be a suitable biomarker of radiation exposure (6). The biodosimetry platform obtained by the experiment could also be used for personalized monitoring of radiotherapy treatments received by patients (12). Gene expression analysis in response to radiation was done in human lymphocytes and peripheral blood leukocytes using three different techniques: microarray, multiplex quantitiative real time PCR (MQRT-PCR) and nCounter Analysis System. A set of genes was found to be suitable for biological dosimetry using peripheral blood. Four of the genes (CDKN1A, GADD45A, PHPT1, and CCNG1) show good agreement between the three methods and the up-regulation of expression in blood and lymphocytes was detected by all the three techniques. These biomarkers could potentially be used for monitoring radiation exposure during radiotherapy and radiological incidents (13).
A novel study was done using blood from patients receiving targeted radiotherapy ( 131 I-mIBG) to characterize biomarkers that may be useful for bio dosimetry. As an alternative biodosimety approach, real time PCR analysis was done for the gene expression and the data showed that transcripts which have already been proven as biomarkers of external exposures in radiotherapy patients are also good early indicators of internal exposure. Three transcripts showed that modulation in gene expression were still significant enough to differentiate between exposed and unexposed samples after 96 hrs of radiopharmaceutical treatment. A bio dosimetry model for gene expression was developed to predict absorbed dose based on modulation of gene transcripts within whole blood. Thus, this biodosimetry for internal radiation dose or the panel of responsive genes obtained from this study could be used for establishing triage in affected areas due to dirty bombs or nuclear reactor accidents at least by rapidly sorting out the 131 I-exposed from unexposed individuals.
Thus, these selected genes could be strong biomarkers of both external and internal exposures to humans (14).  This article focuses on gene expression analysis of Drosophila melanogaster (fruit flies). Compared to humans, biodosimetry information can be obtained in a more controlled manner in animal models because the dose received in humans are usually not known, the exposures may be non uniform and the dose rates may not be known.
Data collection may not be reliable and uniform post irradiation because a lot of variables have to be taken into consideration like age, health, sex, genotype, time since exposure to radiation, personal lifestyle like cigarette smoking, tobacco and alcohol habits (5,17). Drosophila melanogaster is a model organism with a useful lifespan ( ̴ 2 months) and a long history in radiation experiments. Its genome has been sequenced, and many genes in Drosophila are homologous with human genes (18,19). The gene expression results suggest stress, metabolism, reproduction and mitochondrial function as mechanisms involved in the radiation response (20). The data was taken for five radiation doses (plus a control), at 3 time points. The setup of this data allows it to be repurposed for a new analysis that examines the response of genes as a function of radiation dose.
The aim of this study is to secure a set of genes that are responsive to radiation in a predictable way. These genes, particularly if homologous to human genes, have potential uses in radiation dosimetry.

Methodology
The data used in this paper is obtained from data submitted to the gene expression omnibus by Antosh et al (posted under the reference number GSE47999). Normalized data was calculated using the DESeq (21) package in Bioconductor (22).
The data was obtained from an RNA-sequencing gene expression experiment on drosophila melanogaster. at ages 2, 10 and 20 days after irradiating them Flies were irradiated with x-ray exposures of 0, 10, 1000, 5000, 10000 and 20000 roentgen (a 1 roentgen radiation exposure is ≈ 0.01 gray; here we will use the terms "exposure" and "dose" equivalently). The irradiation came inside a chamber containing cesium-137.
Samples were taken at 2, 10 and 20 days after irradiation, with 3 samples per experimental condition (except for sample for 0R, Day 20, where one sample failed quality check). Our re-analysis of this data was done to identify the genes that changed in a predictable way from control, as a function of dose. Genes that behave in a predictable way could potentially be used in a future bio dosimeter. Genes with R 2 >0.9 were selected as behaving linearly. In a secondary analysis, the data for 10 R flies was removed (since this is a very small dose of radiation for fruit flies). The linear analysis described above was run again. In both of these analyses, genes were only selected as linear if at least four radiation doses passed the present/absent cutoff (described in the paragraph immediately before this).
As an additional analysis, gene expression data was examined for "spikes" in fold change. For each gene, at each time point, a set of fold changes was examined (one fold change for each radiation dose). Genes were marked as having a spike if the largest fold change was at least five times greater than the second largest fold change.
Additionally, genes were only counted as having a spike if the fold change of the spike was > 1 (meaning that the average expression at the spike dose was greater than average gene expression in the corresponding control).
For each time point, and for overlaps between time points, genes found to be significant (meaning, linear or spiking) were analyzed as a group using GOStat (23) to see if any biological functions were had a statistically significant amount of genes in the group. Gene ontologies with a corrected p value < 0.05 were selected.
Genes were examined for human homologues using homologene (19) and functional information was found using flybase (18).  FBgn0030189, FBgn0031713, FBgn0032393, FBgn0037020, and FBgn0051864) was found to have a linear response in all time points. Table 1 shows the set of those 6 genes, including homology to human genes (19) and functional information (18) as response to stress, receptor activity, signal transducer activity, detection of bacterium and biotic stimulus and response to DNA damage stimulus. In the genes found in the overlap of day 2 and day 10, overrepresented gene ontologies included peroxisomal transport and NADPH activity. In the genes found in the overlap of day 2 and day 20, overrepresented gene ontologies included several pathways related to the peroxisome, DNA helicase activity, response to hypoxia and telomere maintenance.

Analysis of Linear Behavior with Full Dataset
The overlap between genes in days 10 and 20 found the gene ontology for stress response to be overrepresented. Gene ontologies overrepresented in genes found to be linear at all three time points (2, 10 and 20 days) included peroxisome, DNA helicase activity, ATPase activity and telomere maintenance.

Analysis of Linear Behavior with Lowest Dose Not Included
In the lifespan experiment that accompanied this dataset (20), lifespan effects on fruit flies were not seen until a radiation exposure of 10,000 roentgen (an approximate radiation dose of 100 Gy). The smallest dose in this analysis is 10 roentgen, which is 0.1% of that dose. It is possible that the 10 roentgen dose in this experiment may produce some gene expression at the level of noise. To address that possibility, a secondary analysis for linear behavior was run where the data from 10 roentgen were not included. The results are summarized in figure 2. In this analysis, 13 genes are found to be linear at all three time points. This list includes 5 of the 6 genes found to behave linearly at all three data points when the 10 R data was included in the analysis ( Table 1). The 6 th gene, FBgn0031713, was excluded only because R 2 = 0.88 at day 20. The 13 genes in the overlap are described in Table 2. Of these 13 genes, 4 have human homologues.
A GOStat analysis (23) was run on genes found to be linear at each time point, and also separately on overlaps between time points. Full lists of significant gene ontologies from the analysis can be found in Supplemental Table 2A- included pathways related to peroxisomes, telomere maintenance and Wnt signaling.
No gene ontologies were significantly overrepresented in the overlap between genes linear at day 2 and day 10 postirradiation.

Analysis of Genes for Spikes in Expression
In addition to linear behavior, another potential methodology for using gene expression as a dosimeter would involve genes that "spike"; meaning that a given gene sees a large amount of expression (compared to control flies) at a given radiation dose.
In order to search for such an effect in this dataset, we looked for genes where the fold change was at least 5 times higher at one radiation dose than at any other radiation dose examined. The results are shown in figure 3A, and in Supplemental Tables 3A-G and 4A-G. Zero genes were found in the overlap between all three time points, which suggests that there may be no good candidate genes for a biological dosimeter.
Similar to the linear analysis, we performed the analysis a second time with the data for 10 roentgen radiation exposure removed. Results are shown in Figure 3B, and Supplemental  The analysis with the lowest radiation dose (10 roentgen) did not include the following:  For genes with spikes at day 2 postirradiation, only four genes spiked at doses less than the maximum dose. Two genes spiked at dose 5,000 R; all overrepresented pathways in GOstat were due to FBgn0013745 (similar to the analysis including 10 roentgen). Two genes spiked at dose 10,000 Ryolk protein 1 (as in the analysis including 10 roentgen) and FBgn0013675, which resulted in overrepresented gene ontologies related to oxidative response.
 For genes with spikes at day 10 postirradiation, seven genes spiked at dose 10,000 roentgen. Six of these seven genes were related to reproduction, and include yolk proteins 1, 2 and 3.
 For genes with spikes at day 20 postirradiation, one gene (FBgn0053222) spiked at dose 5000 R (the same gene as the analysis including 10 roentgen).
GOstat results related to the results reported above can be found in Supplemental Tables 5A-H.

Discussion
A radiation dosimeter based on gene expression could result in the better diagnosis of radiation dose in patients, and thus may help in saving lives after a nuclear event or accidental radiation exposure. The results of this paper indicate several candidate genes that have potential to be used for that purpose. In particular, it seems that the best candidates may be the genes listed in Tables 1 and 2 that have human homologues.
One particularly interesting candidate gene is Irbp (inverted Repeat Binding Protein), which was found to behave linearly in all three data points, both with the full data set and with the lowest dose removed. Irbp is related to DNA repair. It is reasonable to predict that DNA damage is linear with radiation dose; thus, it is logical that some DNA repair genes may respond linearly in expression. Irbp has homologues in organisms that are as complex as humans and chimpanzees, and also in organisms such as Japanese rice (19).
Another possibility, based on the application of GOstat results, is to look at particular cellular functions. In particular, the function of protein kinase CK2 may be useful at time points soon after radiation exposure. Protein kinase CK2 was overrepresented in the GOstat analysis for genes found to behave linearly 2 days after irradiation, with a very high statistical significance. Perhaps the functionality of this protein kinase could be measured directly as a function of radiation to produce a different type of radiation dosimeter. Tables 1 and 2 Tables 1 and 2 continue to respond linearly at more times postirradiation, including times < 2 days?

Several genes listed in
 How are these results affected by the energy and type of irradiation?
Further development of this methodology is needed before it can be applied to patients, but these results suggest the possibility of a successful gene expression radiation dosimeter.

Author Disclosure Statement
No competing financial interests exist.

19.
Anonymous ( Day 20 Post-Irradiation at Dose 5000 R. (Analysis with lowest dose discluded.) nanoparticles are a new type of photosensitizer, which generates cytotoxic singlet oxygen molecules upon activation by x-rays. In this paper, we report the use of copper cysteamine nanoparticles, targeted to tumors using pH-Low Insertion Peptide.
In an in vivo study, results show significant tumor destruction under x-ray activation.
An analysis of variance shows that mice treated with targeted particles had a significantly different tumor sizes than mice treated with no particles, as well as mice treated with non-targeted particles. An additional analysis of variance shows that the use of targeted copper-cysteamine nanoparticles affected the survival time after irradiation, compared to irradiation using no particles on mice. This work confirms the effectiveness of Copper-Cysteamine nanoparticles, targeted to tumors, as a photosensitizer when activated by radiation therapy. Combined with radiation therapy, targeted and non-targeted Cu-Cy nanoparticles are good candidates for photodynamic therapy in deeply seated tumors.
Some advantages of PDT are the available options for photosensitizer and therapeutic dose, time of irradiation post treatment and light fluence rate (which can be adjusted to target biological tissues (8)). Although many photosensitizers have been developed, only a few have shown successful results in vitro and in vivo and made it to clinical trials (14). One of the drawbacks of PDT is tissue penetration ability because of the fact that the wavelengths of light for most of the clinically approved photosensitizers are in the UV/visible range. This limits the use of conventional PDT methods to skin (surface) tumors only and are not effective for deep tumors (11,12,(25)(26)(27)(28). Another disadvantage is that the quantum yield of ROS production is lowered under physiological conditions because photosensitizing drugs have poor solubility in water and are easily aggregated due to a hydrophobic nature (3,11,28). Recently, nanomaterials combined with photosensitization drugs have been an important method in photodynamic therapy to overcome the limitation of conventional photosensitization drugs by increasing the cellular uptake and solubility of drugs in water (3).
As stated above, one important issue with photodynamic therapy is the tissue depth at which it can be used as a treatment. One way to address this issue is to use particles that can interact with more energetic photons. Depending on the source of excitation energy, nanoparticles can be designed to be excited by near infrared light (NIR), internal light and X-rays (12). Near infrared light (NIR) can be used to excite upconversion nanoparticles deep in tissue, with higher penetrating capacity compared to visible light and low phototoxicity to normal cells and tissues (26,29). The upconversion nanoparticles showed a strong photodynamic effect on MB49 cells upon irradiation with 980 nm near infrared light (30). However, its penetration ability is still limited compared to X-rays, and it requires high laser light intensity. Further, it is difficult to design and synthesize because the energy gap of near infrared-absorbing photosensitizers is narrow, and the quantum yield of singlet oxygen is usually low (12). Another method for photodynamic therapy is to attach a nanoscintillator to a photosensitizer. When this is done in vivo and exposed to radiation, the nanoparticles emit scintillation. This light is absorbed by the photosensitizers, resulting in the release of singlet oxygen at the tumor site for effective cancer killing. Another alternative strategy is to use luminescent nanoparticles instead of light sources in vivo to support photodynamic therapy with more localized therapy and less potential damage to healthy cells (18,27). X-rays (0.05 -6 MeV) have more tissue penetrating ability than UV/visible/infrared light, which makes it a potential candidate to initiate photodynamic therapy for deeply seated tumors (12). At present, the use of high energy X-rays has been the most common radiation therapy treatment (31). However, radiation therapy often impacts healthy tissue as well as tumors. If the effect of radiation on the tumor can be enhanced, less radiation could be used to get the same effect thus reducing the side effects and damage to the healthy tissues (32). A combination of conventional radiation therapy with photodynamic therapy has been an exciting technique for deep tumor penetration (33) and has the potential to result in lower doses of radiation when scintillation nanoparticles are attached to photosensitizers (12,18). The nanoparticles emit light when induced by ionizing radiation; the scintillation activates the photosensitizers and results in the release of singlet oxygen. In this case, photodynamic therapy takes place even without the aid of an external light source, and the effectiveness of the radiation is increased (18). Since the site of damage from photodynamic therapy depends on the location of the photosensitizer at the time of irradiation, (20) conjugating the particles with tumor specific targeting molecules can enhance the uptake of particles with efficient cancer treatment reducing the damage to the healthy tissues and important organs near the tumor with the reduced radiation dose (32). X-rays can initiate the photodynamic agent (LaF 3 : Tb 3+ -meso-tetra ( 4-carbosyphenyl) porphine (MTCP)) scintillating nanoparticle, even at low dose, for deep cancer treatment (34). The core of a nanoscintillator coated with a mesoporous silica forms an integrated nanosystem which when irradiated by X-rays (25).  showed enhanced X-ray damage by gold nanoparticles treated with a new synthesis method of polyethylene glycol modification. Trifluorocerium-verteporfin (CeF 3 -VP) conjugates, lanthanide complexes, Copper and cobalt co-doped zinc sulfide (ZnS:Cu,Co) afterglow NPs, and nanoscintillator coupled porphyrins have been shown to produce singlet oxygen when activated by X-ray and are effective for cancer cell destruction (4,5,35,36). Zhang et al. used a conjugated semiconductor scintillator particle as a photosensitizer with ionizing radiation, and found diminished oxygen dependence (37). The combined effect of radiation therapy and photodynamic therapy with indocyanine green as a sensitizer resulted in killing of MCF7 human breast cancer cells with a reduction in percentage cell viability, down to 3.42%. A one way ANOVA was used to analyze data for statistical differences (p < 0.05) (38). When activated by X-rays at 90 kV, energy was transferred from Ce 3+ -doped lanthanum(iii) fluoride (LaF 3 :Ce 3+ )/DMSO nanoparticles to protoporphyrin IX (PPIX) with the production of singlet oxygen to kill cancer cells (39). Porphyrin conjugated with SiC/SiO x nanowires has been an efficient source of singlet oxygen at low doses of 6MV X-rays (0.4 -2 Gy), showing the enhancement of radiation therapy for cancer treatment (40).
In addition to improving upon the depth that photodynamic therapy can reach, another opportunity for improvement is the use of active targeting agents like peptides, antibodies and proteins. These agents could reduce the side effects to the surrounding healthy tissues (3,41) and problems associated with multidrug resistance (11). There is a need for more precise photosensitization drug delivery into target cells and tissues (3), and efforts have been made to search for alternative photodynamic therapy methods for deep tissue penetration (26,27). Targeted photodynamic therapy has been a new promising therapeutic strategy that enhances specificity and efficiency of photodynamic therapy by improving the delivery of photosensitizers to cancer tissue (15). Yoon et al. successfully inhibited tumor growth using the hydrophobic photosensitizer chlorin e6 (Ce6), conjugated with tumor targeting hyaluronic acid nanoparticles (HANPs), to generate singlet oxygen in tumor cells when irradiated by laser. They analyzed the differences between experimental and control groups using a one-way analysis of variance (ANOVA) and found their results to be considered statistically significant if p<0.05 (42).
Copper cysteamine nanoparticles (Cu-Cy, Cu 3 cl(SR) 2 ) are a new option for photosensitization and radiation therapy. They were used to kill SW620 colorectal cancer cells by inducing apoptosis as well as autophagy. The difference between the control and experimental groups was determined using Student's t test and a one oneway analysis of variance (ANOVA) (43). Copper cysteamine particles under X-ray activation generated singlet oxygen ( 1 O 2 ) and were successful at killing MCF-7 cells both in vitro and in vivo and can be used in the treatment of both shallow and deep cancers (44,45). Copper cysteamine has been demonstrated as an X-ray activated nanoparticle in photodynamic therapy for cancer treatment, which when conjugated by tumor specific targeting molecules can enhance the uptake (44).
In this paper, we demonstrate the use of copper cysteamine nanoparticles to enhance radiation therapy, using photodynamic therapy. We targeted copper cysteamine nanoparticles to tumors using the targeting peptide pH-Low Insertion Peptide (pHLIP), which targets molecules to tumors using the property that tumors have low pH. Among many uses, pHLIP has been used to effectively target gold nanoparticles to tumors and to treat cancer using gold nanoparticles (46)(47)(48)(49).

Materials and methods
Preparation of pHLIP conjugated copper cysteamine nanoparticles 2 mg of Var3 pHLIP (Ala-28-Gly), from CS Bio Company, was added in 5 mL of deionized water followed by the addition of 3.19 mg of 1 Ethyl-3-(3dimethylaminopropyl) carbodiimide (EDC) under mild stirring for 10 minutes at room temperature. After adjusting the pH to 7.5 using NaoH, 5 mL of 1 mM copper cysteamine water solution was added under constant stirring overnight at room temperature in a dark environment. The copper cysteamine-pHLIP conjugates were centrifuged at 4400 rpm for 25 minutes and washed with deionized water 3-4 times.

Cell Culture
JC Breast murine cancer cells of BALB/cRos strain were purchased from American Type Culture Collection (ATCC) and were grown in Roswell Park Memorial Institute (RPMI) medium with L-glutamine and sodium bicarbonate, 10% Fetal Bovine Serum (Sigma-Aldrich) and 0.1% Ciprofloxacin. The cells were maintained in a humidified atmosphere at 5% carbon dioxide at 37 degrees centigrade in an incubator.

Animal Models and Cell Injection
All animal work followed the guidelines of URI IACUC protocol AN1516-003. Males and Females, 18-25g Balb/c AnNHsd, 3-4 weeks mice were ordered from Envigo.
This strain of mice came to Harlan Sprague Dawley (Hsd-now Envigo) from National Institutes of Health, Bethesda. The NIH received this strain from Andevont (An). 1.5 million cells suspended in 100uL RPMI were injected subcutaneously on the right flank of the mice using 1 mL 27 G 1/2 latex free BD syringes.

Radiation Therapy on Mice
Mice were divided into 6 treatment groups: i) targeted copper cysteamine + radiation ii) untargeted copper cysteamine + radiation iii) PBS (control) + radiation iv) targeted copper cysteamine v) untargeted copper cysteamine vi) PBS (control). In total, 51 mice (24 males and 27 females) were used for the experiment.
Treatment was undertaken when the tumor size reached approximately 4-8mm. Mice were anesthetized using isoflurane gas. For groups of mice given nanoparticles, the particles were injected intratumorally in 20 μL PBS at a particle concentration of 0.8 μg/μL. For the groups given radiation therapy, the mice were irradiated 30 minutes post injection of particles at an irradiation dose of 5 Gy. No external X-ray filter was used, and the source to surface distance was set to 30.5 cm with a field size of 18.3 by 20.1 cm. The current and voltage settings of the X-ray machine (a Faxitron MultiRad 350) were 90 kVp and 30 mA. The non-irradiated mice were placed in the x-ray chamber in the same settings but with no irradiation. The tumor size was measured daily using digital Vernier calipers (VCD001, from United Scientific Supply) to get the tumor volume. The tumor volume was calculated using the formula: tumor volume = ½ length * width 2 (50). Mice were euthanized if they reached the endpoint size of 20 mm, or if they showed signs of distress.

Particle characterization
The coppercysteamine nanoparticles were synthesized in The University of Texas at Arlington in Wei Chen's lab along with the singlet oxygen measurement and photoluminescence and X-ray luminescence measurements (44).

Statistics and Analysis of Data
In total, 51 mice were used -24 males and 27 females. Each of the radiation therapy groups had 3 males and 4 females, whereas the non-RT groups had 5 males and 5 females.
The effect of experimental variables on tumor size in our experiment was quantified by running an analysis of variance (ANOVA), using the command "anova" in the statistical computer language R (51). The input given was a linear model (lm command in R). The independent variable in the linear model was tumor volume, and the dependent variables were treatment type, time of measurement after irradiation, age at irradiation, radiation dose (0 or 5 Gy), sex (M or F) and original tumor volume.
Treatment types were run two at a time to generate a comparison between the following pairs of treatments: i) targeted copper cysteamine particles and untargeted copper cysteamine particles; ii) targeted copper cysteamine particles and no particles; iii) untargeted copper cysteamine particles and no particles. p values for individual variables, as well as interactions of variables were determined using the F test (part of the anova command in R). p values were ruled significant if the Bonferroni correction criteria was met. Including interactions, 59 p values were found for each pairwise comparison of treatment types. We used 0.05/59 = 0.000847 as the cutoff P value for statistical significance. An analysis of variance (ANOVA) was performed to compare the effects of (targeted copper cysteamine particles + radiation) with the effects of (non-targeted copper cysteamine particles + radiation). As described in detail in the methods section, the dependent variable of the analysis was tumor size and the independent variables were time after irradiation, sex (male or female), radiation dose (0 or 5 Gy), treatment type (targeted particles or non-targeted particles), age of mouse at irradiation, and volume of tumor at time of irradiation. Mice with a treatment of no particles were excluded from this analysis, so that the variable for treatment type would be a 2-factor comparison of targeted and non-targeted particles. Table 1 shows the ANOVA results for an analysis where only two treatment types were included: targeted and non-targeted particles. In this analysis, the p value for treatment type is significant (less than 0.000847, the Bonferroni cutoff). This indicates that mice treated with targeted particles had a significantly different tumor size than mice treated with non-targeted particles, even when other relevant experimental variables were also considered.

Tumor Size Data and Analysis of Variance
In addition to treatment type, the p values for the following variables were also significant: time after irradiation, radiation dose, treatment type, age of mouse at irradiation, volume of tumor at time of irradiation. The p value for sex of the mice shows that sex does not play an important role in the experimental outcome. Several interaction terms were also significant in the analysis of variance. Of particular note are: the interactions between time after irradiation and radiation dose (p value < 2.2*10 -16 ), radiation dose and treatment type (p < 5.02*10 -9 ), time after irradiation and age at irradiation (p<2.2*10 -16 ), treatment type and age at irradiation (p< 3.576*10 -11 ), time after irradiation and tumor volume at irradiation (p< 6.739*10 -10 ), age at irradiation and tumor volume at irradiation (p<0.000219). See supplemental table 1 for full information on interaction terms for this analysis. Table 2 shows the ANOVA results when the following treatment types were included: non-targeted particles and no particles. In this analysis, all experimental variables tested had a significant effect on tumor size, including sex. Notable significant interaction terms included: time after irradiation and radiation dose (p<2.2*10 -16 ), time after irradiation and age at irradiation (p<2.2*10 -16 ), radiation dose and age at irradiation (p<2.308*10 -13 ), time after irradiation and tumor volume at irradiation (p< 2.2*10 -16 ), radiation dose and tumor volume at irradiation (p< 6.238*10 -6 ), treatment type and tumor volume at irradiation (p<1.001*10 -6 ), age at irradiation and tumor volume at irradiation (p<0.0029175). See supplemental table 2 for full information on interaction terms for this analysis. Table 3 shows the analysis of variance when the following treatment types were included: targeted particles and no particles. Similar to table 1, all experimental variables included were significant except sex. Notable significant interaction terms included: time after irradiation and sex (p<5.753*10 -5 ), time after irradiation and radiation dose (p<2.2*10 -16 ), radiation dose and treatment type (p<1.15*10 -8 ), time after irradiation and age at irradiation (p<2.2*10 -16 ), treatment type and age at irradiation (p<1.243*10 -10 ), time after irradiation and tumor volume at irradiation (p<2.2*10 -16 ), radiation dose and tumor volume at irradiation (p<0.002), age at irradiation and tumor volume at irradiation (p<0.003374). See supplemental table 3 for full information on interaction terms for this analysis.

ANOVA Analysis of Survival
The number of mice used in this experiment was insufficient to run an effective log rank test for differences between survival curves. As a substitute analysis, an analysis of variance was run with time between irradiation and death as the dependent variable.
(Here, death is defined as either actual death or as reaching a humane endpoint following our institution's IACUC policies.) The independent variables were sex (M or F), radiation dose (0 or 5 Gy), treatment type (targeted particles, non-targeted particles, no particles), age of mouse at time of irradiation, and size of the mouse's tumor at time of irradiation. Including interaction terms, there were 29 variable combinations assessed for significance; we used a Bonferroni cutoff p value of 0.05/29 = 0.00172 to claim significance.
A comparison using mice treated with targeted particles or no particles found a statistically significant effect from treatment (targeted particles versus no particles), sex and radiation dose. Thus, the anova analysis indicates that targeted particles increase survival time, compared to irradiation using no particles. Notable interaction terms that were significant included radiation dose with treatment type, and radiation dose with tumor size at time of irradiation. The input and output of this analysis is supplemental tables 4 and 5. Analyses run with the other combinations of treatments (targeted particles versus non-targeted particles, non-targeted particles versus no particles) found no significant effects from the treatment differences.

Discussion
In this paper, we demonstrated that copper cysteamine nanoparticles, targeted with pH-Low Insertion Peptide, can be used to reduce tumor size and to increase survival in mammals. Copper cysteamine can be used in the treatment of both shallow and deep tumors because it can be activated by X-rays as well as light (44). In this result, we particularly emphasize the effect on tumor size.
The targeted Copper cysteamine particles showed the enhanced radiation effect with better tumor killing in both the male and female mice. The sex of the mice might not be an important variable in this kind of experiment as this parameter was not statistically significant in targeted vs non tartgeted and targeted versus no particles.
However for nontargeted versus no particles it played an important role. The original volume of tumor and age factor of mice at irradiation time are also very important to be considered while performing the experiment. The dose of the radiation given to the mice could be altered to see the affect in the tumor size as the radiation dose is statistically significant.
One particularly important variable that was not tested in this paper is radiation energy. Few, if any, photoluminescent particles have been shown to work at energies as high as 90 kVp, as shown in this paper. However, most clinically relevant energies are higher still.
Overall, this paper represents a firm demonstration of the effectiveness of coppercysteamine nanoparticles in the treatment of mammalian cancer.