Radiation therapy plays an increasingly important role in the treatment of hepatocellular carcinoma (HCC), particularly in the management of locally advanced unresectable HCC [1, 2]. However, radiation-induced liver disease and radiation resistance are major factors that limit the success of radiation therapy [3–5]. To improve the effectiveness of radiation therapy, image-guided radiation that provides real-time imaging of the tumor target during treatment has been study. Real-time imaging can help compensate for normal movement of the internal organs from breathing and for changes in tumor size during treatment. Moreover, in contrast to tumor control it enables full therapeutic efficacy while reducing the degree and frequency of invasive interventions [6].

On the other hand, radioprotectors and radiosensitizers, chemicals that modify a cell’s response to radiation are also under development. Radioprotectors are drugs that protect normal cells from damage caused by radiation therapy [7]. In contrast, radiosensitizers make tumor cells more susceptible to radiation damage and can increase the damaging effects of radiation while minimizing exposure to normal and healthy cells [8]. Several radiosensitizers, such as misonidazole, metronidazole, tirapazamine, trans sodium crocetinate are under study [9]. In addition, some anticancer drugs, such as 5-fluorouracil and cisplatin, can make cancer cells more sensitive to radiation therapy as well [10–12]. Together, radiosensitizers could enable radiation therapy at a lower dose, but would not affect treatment efficiency.

Nanoparticles have played a key role in the enhancement of the radiation therapy by acting as both a therapeutic and a carrier for other therapeutics. Particularly, inert and biocompatible gold nanoparticles (GNPs) with tunable size and low osmolality hold many potential advantages [13]. Gold absorbs approximately 3-times more than iodine at 20 and 100 keV; the dose enhancement factor for gold is adjustable depending on the beam’s energy and the amount of gold delivered; the biodistribution of GNPs can be imaged in real time before a therapeutic dose is delivered and used for treatment planning and quantified prediction of dose enhancement; moreover, GNPs allow surface immobilization of antibodies, peptides, aptamers or drugs. Thus, in recent years, using GNPs to enhance the radiation absorbed dose deposited in tumors from X-ray has gained increasing attention [14–17]. Ligands functionalized GNPs could enhance radiation sensitivity in radiation-resistant human prostate cancer cells at keV and MeV X-rays [15]. Chithrani et al. also showed a range of cell line specific responses including decreased clonogenic survival, increased apoptosis and induction of DNA damage could be induced when exposure to 1.9 nm GNPs and X-rays [17].

Ligands against cancer cell surface biomarkers or extracellular matrix proteins have led to the applications of ligands for specific delivery of therapeutic agents [18]. Asialoglycoprotein receptor (ASGPR) is a hepatocyte-specific receptor, and approximately 5 × 10⁵ ASGPRs are expressed on the sinusoidal surface of a single liver cell [19]. ASGPR mediates the capture and endocytosis of galactose- or N-acetylgalactosamine-terminating glycoproteins [19, 20]. Thus, galactose, asialofetuin, acetylgalactosamine and asialoorosomucoid targeting ASGPR are commonly used as homing agents for drug delivery [21].

We propose that ASGPR-targeted GNPs may enhance the cytotoxic effects of radiation therapy, as well as concentrate the effect on targeted tumor cells in tissue. β-d-galactose (GAL) as a homing agent thus is immobilized onto bare GNPs surfaces. Further, to prevent GNPs from being quickly phagocytosed by reticuloendothelial system (RES) [22], biodegradable and biocompatible polyethylene glycol (PEG) is co-grafted to form a novel composite known as GAL-PEG-GNPs (Fig. 1). Results showed that compared to bare GNPs, GAL-PEG-GNPs can be more quickly and effectively uptaken by ASGPR over-expressed HepG2 cells. HepG2 cells treated by GAL-PEG-GNPS showed significant DNA double-strand breaks and cell apoptosis after 6-MeV X-rays irradiation. These results demonstrated that our GAL-PEG-GNPs hold potential as a radiosensitizer for HCC therapy.

Nanoparticles can permeate leaky angiogenic endothelium providing some tumor specificity. Additionally, blood clearance of nanoparticles is slower than small molecules, such as iodine contrast media, and thus nanoparticles can stay in the blood for hours [23]. Moreover, drugs can be bound to the nanoparticles or be entrapped inside the nanoparticles, allowing flexible and multifunctional therapy [24]. Nevertheless, nanoparticles have the potential to revolutionize cancer diagnosis and therapy [25]. Among various nanomaterials, particularly GNPs are very attractive. They can be safely administered with minimal inflammatory activation and few side effects. GNPs are very convenient to realize a variety of surface functionalization with various biological molecules to help improve stability, drug-carrying, and tumor-targeting [12]. Furthermore, gold allows high absorption and enhancement of ionizing radiation, as well as superior X-ray attenuation for biomedical imaging application. Moreover, other physical properties including surface plasmon resonance and Raman scattering activity have been exploited in fluorescent imaging, photothermal therapy, and many other applications [26, 27]. At KeV or MeV radiation energies, GNPs can selectively improve radiosensitivity of tumor cells leading to increased cell killing and have been trialed on various cancer cell lines and animals [28–31]. Theoretically, the dose enhancement achieved by GNPs could be more than 200 % [16]. Though the mechanisms are still unclear, radiosensitization is generally attributed to increasing photo absorption of high-Z elements, and the resulting transfer of a larger portion of primary ionizing photon energy to tumor tissue.

Our previous results showed that bare GNPs could improve the radiosensitivity of HepG2 cells, elevate DNA damage levels, and induce cancer cells apoptosis in vitro [14]. To deliver a high dose of GNPs directly to malignant cells, and to specifically radiosensitize them while minimizing side-effects, we further graft GAL targeting ASGPR for delivering GNPs to HepG2 cells in vitro. First of all, GNPs were prepared by sodium citrate reduction followed by surface functionalization with SH-PEG-NH₂ molecules and surface deactivation using SH-PEG molecules. The grafted amine groups were used to covalently immobilize GAL.

The size of GNPs for liver cancer therapy should not exceed 200 nm, since the size of liver sinusoidal fenestrations have 200 nm effective size limit. On the other hand, GNPs may experience rapid renal clearance if they are smaller than 10 nm in diameter. Size is also an influencing radiation sensitivity parameter, and large sized GNPs have superior dose enhancement factor [32, 33]. Bergen et al. tested the efficacy of GNP size ranging from 50 to 150 nm and concluded that GNPs in 50-nm diameter could have the highest radio enhancement factor (1.43 at 220 keV) and also have the highest cellular uptake [16]. However, a recent report showed that GNPs in 18-nm diameter have more cell internalization [34]. Though the optimal size of GNPs is inconclusive so far, there is a tradeoff between effective dose enhancement of large GNPs and the effective clearance of small ones. This remains a hurdle for their application in cancer imaging and therapy. The average size of our GNPs and GAL-PEG-GNPs is 23 and 34 nm, respectively. Therefore, our nanoparticles prepared in this study are in principle suitable, in terms of size for radiosensitization applications, exploiting the enhanced permeability and retention effect and avoiding renal clearance. The larger size and peak shift of GAL-PEG-GNPs are attributed to surface PEGylation and GAL grafting. PEGylation can further reduce the reticuloendothelial system uptake and thus significantly increase circulation time. In addition, PEGylation decreases aggregation of GNPs owing to passivated surfaces and increase solubility in buffer and serum due to hydrophilic ethylene glycol repeats. It was reported that PEGylated nanoparticles show higher tumor accumulation versus background.

Both GNPs and GAL-PEG-GNPs at various concentrations showed cytotoxicity, and their IC₅₀ values are 5.001 and 4.997 μg/mL, respectively. Compared to the published data of GNPs cytotoxicity [35], our GNPs are relatively less cytotoxic. In this study, we merely used 1/5 concentration (1.0 μg/mL) to minimize the cytotoxicity on HepG2 cells, and thus such low level of cytotoxicity can be neglected.

ICP-MS analysis revealed that the amount of GAL-PEG-GNPs endocytosed by HepG2 cells was approximately three times more than that of GNPs dose, suggesting GAL significantly increased the gold uptake. Although both internalizing and noninternalizing epitopes on cell membrane can be targeted, if ligands bind to noninternalizing ones, drugs or nanoparticles may accumulate around the cells and enter into cells either by passive diffusion or normal transport mechanisms. In contrast, when GAL bind to internalizing epitopes, like ASGPR, the binding triggers ASGPR-mediated uptake and hence GAL-PEG-GNPs uptake is thus more effective. TEM images showed that GNPs were mainly distributed in mitochondrion. Few reports have demonstrated a role for reactive oxygen free radicals (ROS) or the involvement of mitochondria as mechanism of GNPs radiosensitization [36, 37]. In this study, we confirmed that elevated levels of DNA damage which may be a direct result of impaired mitochondrial function manifested by increased oxidation and loss of membrane potential, resulting in HepG2 cells proliferation inhibition. It was further confirmed by clonogenic analysis. After exposure of cells to various radiation doses (1–8 Gy zone), clone quantities in GAL-PEG-GNPs-treated group were significantly less than that in GNPs-treated and control groups mainly due to more amount of gold uptake.

The cell cycle analysis showed that cells were propelled to enter G2/M phase. GAL-PEG-GNPs induced approximately 27 % of cells in the G2/M phase; the percentage was more than that of GNPs induced. G2/M phase is the most radiosensitive phase of a cell cycle, and thus more accumulation in G2/M phase can enhance the radiation sensitivity. It was reported that GNPs can trigger activation of the CDK kinases, leading cancer cells to accumulate in the G2/M phase [37, 38]. Additionally, p53 protein may mediated G1 arrest and p53-independent G2/M phase arrest over 2–10 h after radiation [39]. Moreover, cell cycle kinetics can be changed by GNPs after radiation. Previous reports demonstrated that GNP radiosensitization can increase sub G1 population or accumulate cells in G2/M. Additionally, surface functionalization of GNP using a range of different biological moieties not only enables them maintain stability and bio-compatibility but also leads to accumulation of GNPs in tumors compared to healthy tissues [40]. Thus, it is likely to provide significant therapeutic advantage and may potential benefit personalized medicine.

γ-H2AX assay results showed a significantly higher incidence of DNA-DSBs in GAL-PEG-GNPs-treated group as compared to that in GNPs-treated or control group. Decreased clonogenic survival of HepG2 cells and increased DNA-DSBs demonstrated that GAL-PEG-GNPs achieved better radiotherapy sensitization effect compared to bare GNPs.

Significantly upregulated expression of Cytochrome C, Bax, Caspase-3, and Caspase-9 and downregulated anti-apoptotic protein Bcl-2 were found in GAL-PEG-GNPs-treated cells after radiation. We found antioxidant enzymes including CAT, SOD, and GSH were significantly reduced indicating the increase of free radicals in HepG2 cells. The release of ROS can cause mitochondrial membrane depolarization and oxidization, and subsequently attack intercellular macromolecules, such as DNA, RNA, proteins, leading to above mentioned changes [31, 32]. The cytochrome C, which was released from mitochondria to cytoplasm activated Caspase-9; Caspases-9 further activated Caspase-3 and others [33, 34]. Caspase-3 caused chromatin pyknosis, DNA fragmentation, and DNA fracture leading to cell apoptosis [35]. In contrast, the mechanisms of upregulated Bax and downregulated Bcl-2 through GNPs-mediated radiation are not clear yet, and thus should be further investigated in the future.

In this study, we successfully prepared GAL-PEG-GNPs as a novel radiosensitizer,found our novel nanoparticles could significantly improve the efficacy of radiotherapy in vitro, and briefly discussed the mechanism of radiosensitization. Given the wide range of methods available to further modify these nanoparticles, in ongoing work we are also exploring the use of cell-specific internalizing aptamers (i.e. anti-EGFR aptamers or anti-EpCAM aptamers) grafted PEG-GNPs that can more effectively be taken up by cancer cells overexpressing a cognate receptor, and the dual delivery of chemotherapeutic agents and GNPs to achieve a synergistic therapeutic outcome in vivo. These vehicles may be able to circumvent biological barriers, such as the blood–brain barrier, thus resulting in efficient localization of various agents at the target sites.