A combination hepatoma-targeted therapy based on nanotechnology: pHRE-Egr1-HSV-TK/131I-antiAFPMcAb-GCV/MFH

Combination targeted therapy is a promising cancer therapeutic strategy. Here, using PEI-Mn0.5Zn0.5Fe2O4 nanoparticles (PEI-MZF-NPs) as magnetic media for MFH (magnetic fluid hyperthermia) and gene transfer vector for gene-therapy, a combined therapy, pHRE-Egr1-HSV-TK/131I-antiAFPMcAb-GCV/MFH, for hepatoma is developed. AntiAFPMcAb (Monoclonal antibody AFP) is exploited for targeting. The plasmids pHRE-Egr1-HSV-TK are achieved by incorporation of pEgr1-HSV-TK and pHRE-Egr1-EGFP. Restriction enzyme digestion and PCR confirm the recombinant plasmids pHRE-Egr1-HSV-TK are successfully constructed. After exposure to the magnetic field, PEI-MZF-NPs/pHRE-Egr1-EGFP fluid is warmed rapidly and then the temperature is maintained at 43 °C or so, which is quite appropriate for cancer treatment. The gene expression reaches the peak when treated with 200 μCi 131I for 24 hours, indicating that the dose of 200 μCi might be the optimal dose for irradiation and 24 h irradiation later is the best time to initiate MFH. The in vitro and in vivo experiments demonstrate that pHRE-Egr1-HSV-TK/131I-antiAFPMcAb-GCV/MFH can greatly suppress hepatic tumor cell proliferation and induce cell apoptosis and necrosis and effectively inhibit the tumor growth, much better than any monotherapy does alone. Furthermore, the combination therapy has few or no adverse effects. It might be applicable as a strategy to treat hepatic cancer.

Gene therapy is one of important biological treatments, and suicide genes especially show a good application prospect in the field of cancer treatment 16 . In suicide gene therapies, herpes simplex virus type thymidine kinase (HSV-TK) gene is most commonly used. It can express thymidine kinase to transform non-toxic prodrug ganciclovir (GCV) into toxic GCV-TP to kill tumor cells. A large number of studies have proved the anticancer effects of HSV-TK/GCV system, and better curative effects can be achieved when it is combined with radiotherapy 17,18 . However, two crucial problems must be solved to accomplish comprehensive gene therapy against cancer for a desired curative effect. One is how to ensure safe gene delivery into cells with high transfection efficacy. The other is how to make gene express efficiently and controllably.
Honestly, radiation-gene therapy shows a promising anticancer prospect. The promoter of Egr1, a transcription factor regulating early cell growth, can induce the expression of its downstream genes after ionizing radiation, thereby attaining a spatio-temporal regulation on the target gene expression. Notable achievements in radiation-gene therapy have been obtained by exploiting the radiosensitivity of Egr1 promoter [19][20][21] . In addition, hypoxia response element (HRE), an enhancer sensitive to hypoxia, can promote its downstream genes to express in hypoxia environment [22][23][24] . Shibata et al. 25 has confirmed that five copies of HRE connected with a promoter can increase its downstream gene expression by 500 times in hypoxia condition. In our previous studies, we obtained anoxic radiation double sensitive promoter HRE/Egr1 by coupling Egr1 promoter with HRE and constructed pHRE-Egr1-EGFP plasmids. The resulting HRE/Egr1 could effectively induce and improve EGFP expression in cancer anoxic microenvironment when radiated and the expression level of target gene induced by anoxia and radiation was significantly higher than by radiation alone 17 . However, this method has yet to solve the common problem how to safely and effectively deliver genes into cells in gene therapy.
Two of the major gene delivery vectors, viral vector systems and non-viral vector systems, are currently being used, but both have their own pros and cons. Although viral vectors have been proved efficacious, their small gene capacity, poor target specificity, self-immunogenicity, and especially serious biosafety concerns limit their further application. Despite avoiding the major security risks, non-viral vectors such as lipofection, polyamidoamine 26 , poly(ethyleneimine) 27 , and poly(amino acid) 28 are greatly inferior to viral vectors in transfection efficiency, and meaningful expression of target gene is hardly available in this system. Therefore, the gene delivery remains a challenge because of the lack of suitable vectors.
It is encouragingly that nanocarriers are emerging as a kind of powerful delivery tool in gene therapy. Such vectors retain the advantages of viral and non-viral carriers and remove their disadvantages and therefore become a new promising carrier system. Compared to traditional carriers, nanocarriers have much superiority, including slowly released the genes, maintained effective concentration of the product, improved transfection efficiency and the product bioavailability. Also nanocarriers have no immunogenicity, no genetic toxicity, no cytotoxicity, and no cell transformation or death [29][30][31][32][33][34][35][36][37][38] . Particularly, except the general properties of nanoparticles, super-paramagnetic magnetic nanoparticle gene transfer vectors can produce highly efficient transfection and do directional movement in an external magnetic field, in turn which help to implement targeting gene therapy. Moreover, since magnetic nanoparticles can generate heat by magnetic induction in an external magnetic field, they also can be used for cancer thermotherapy [39][40][41] .
Representing an alternative treatment for cancer, thermotherapy not only can play an anticancer role itself, but also can improve the sensitivity of chemotherapy and radiotherapy [42][43][44][45] . It is investigated that when heated for 30 min at 43 °C, the cytotoxicity of paclitaxel (a chemotherapy drug) can increase by 10-100 times, and the killing ability of some chemotherapeutics with little cytotoxicity at normal temperature can double 46,47 . Wang etc. 48 used radiotherapy combined with thermotherapy to treat advanced lung cancer. The effective rate reached 70% (21/30), while the single-radiotherapy was only 44% (14/32). However, there is an intractable technical difficulty in heating tumor tissue evenly at a desired temperature, with no damage to normal tissue during thermal treatment.
Inspiringly, Jordan, et al. 49 developed magnetic fluid hyperthermia (MFH), a new tumor thermotherapy by combining magnetic induction heating with nanotechnology. This therapy has a targeted position function. In other words, in an external magnetic field, the temperature can specifically rise in the tumor tissue containing magnetic nanoparticles, whereas normal tissue without magnetic particles is not subject to thermal damage. It therefore has a great specificity for cancer treatment. We have previously prepared thermo-sensitive Mn-Zn ferrite magnetic nanoparticles (Mn 0.5 Zn 0.5 Fe 2 O 4 , MZF-NPs) 17,18 . Both in vitro and in vivo experiments confirmed their excellent magnetic responsibility, good heating and thermostatic ability, and good biocompatibility [50][51][52][53][54] . One of the best performances for MZF-NPs is that they can get a set Curie temperature by adjusting the concentration of their magnetic fluid. To be specific, MZF-NPs are strong magnetic materials below the Curie temperature, which can absorb electromagnetic waves to warm up in an alternating magnetic field. Upon reaching the Curie temperature, they will change into non-magnetic materials and lose the ability to absorb electromagnetic waves, so their temperature starts to decrease. After the temperature falling below the Curie temperature, the materials begin magnetic heating again. So cyclically, the temperature always fluctuates at the set Curie temperature, such as 42-44 °C, which is an effective treatment temperature for tumors, inflicting no damage to normal tissue. It thus successfully resolves the problem of temperature controllability in hyperpyrexia treatment, improving the stability and security of thermotherapy 17,55 . Magnetic induction heating by using Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles as magnetic media treated liver cancer and cervical cancer and yielded good outcomes, especially presenting an obvious synergistic effect when combined with As 2 O 3 56,57 . Meanwhile, MZF-NPs modified with PEI have good loading capability, transfect capability and protection ability to DNA and can be used as gene-transfer vectors in gene therapy 17,18,52 .
In addition to therapeutic effect, good targetability is also crucial for tumor treatment [58][59][60] . Especially, the progress of radionuclide therapy largely relies on the development of nuclide targeting system. To select an ideal vector with high specificity, adhesion, penetration and carrying capacity to selectively deliver radionuclide into tumor is the key to improve nuclide tumor uptake and T/NT (the ratio of radionuclide in tumor versus in Scientific RepoRts | 6:33524 | DOI: 10.1038/srep33524 nontumor), and this is also considered as a precondition for nuclide therapy to obtain good curative effects, with an exception of radioiodine to treat thyroid.
As a class of specific tumor antigen in the cell membrane or cytoplasm, α -fetoprotein (AFP) is positive in over 70% of primary hepatic carcinomas, but negative in normal liver or other tissues [61][62][63] . Consequently, it is a good potential antigen for hepatoma-targeted treatment. Owing to its high specificity and affinity in liver cancer cells, monoclonal antibody AFP (antiAFPMcAb) can be used to carry various "warheads" such as chemotherapy agent, radioactive nuclide or toxin to selectively attack AFP-positive cancer cells 64,65 . In our previous study, we prepared radionuclide immune albumin nanospheres ( 131 I-antiAFPMcAb-GCV-BSA-NPs, short for 131 I-antiAFPMcAb-GCV). In these nanospheres, GCV was encapsulated in bovine serum albumin (BSA) nanoparticles. AntiAFPMcAb labeled with radioactive 131 I was grafted to GCV-BSA nanoparticles to enhance the targetablity of 131 I and GCV to tumors with over-expressed AFP. The in vitro and in vivo targeting experiments showed good selectivity of 131 I and GCV for AFP-positive tumors 66 .
In the present study, we developed a combined therapy, pHRE-Egr1-HSV-TK/ 131 I-antiAFPMcAb-GCV/ MFH, for hepatoma by using PEI-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles (PEI-MZF-NPs) as magnetic media for MFH and gene transfer vector for gene-therapy, and investigated the therapeutic effects of pHRE-Egr1-HSV-TK/ 13 1 I-antiAFPMcAb-GCV/MFH on hepatoma in vitro and in vivo, as well as the underlying safety. As shown in the schematic of the combined therapy (Supplementary Figure 1), since PEI-Mn 0.5 Zn 0.5 Fe 2 O 4 nanoparticles have good magnetic property and magnetic induction thermogenesis in an alternating magnetic field, they can be used for hyperthermia therapy. While killing hepatoma cells, 131 I can activate Egr1 promotor to induce HSV-TK gene expression and the expression can be especially enhanced by HRE in hypoxic solid cancer, so the gene therapy can be initiated. The antiAFPMcAb confers the therapy targetability. As a result, a multiple targeting killing of genes, radionuclide and hyperpyrexia against hepatoma can be achieved.

Results and Discussion
Characteristics of MZF-NPs. The resulting MZF was approximately spherical, about 15-20 nm in diameter and some agglomerate, as revealed by TEM (Fig. 1A). To alleviate the agglomeration of MZF-NPs and make them play roles of gene transfer vector and magnetic induction medium, we modified MZF-NPs with PEI, a surface modification agent, whose monomer (-CH-CH 2 -NH 2 -) has a good ability to bind DNA and adhere to cells 67 . Figure 1B is the infrared spectrum analysis of coated and uncoated MZF-NPs, showing some small characteristic peaks of -NH 2 and -CH 2 -at 876 cm −1 , 1316 cm −1 , 1640 cm −1 and 2830 cm −1 , namely the shear vibration peaks of -NH 2 of PEI and asymmetric and symmetric stretching vibration characteristic peaks of -CH 2 -, which indicates that MZF-NPs are successfully coated by PEI.

Identification of pHRE-Egr1-HSV-TK by restriction enzyme digestion.
In the constructed pHRE-Egr1-HSV-TK plasmids, there were two BglII restriction sites and the fragment between the two sites contained HRE about 380 bp. The fragment between the two sites on MluI and NheI was Egr1 and the length was about 1152 bp. HSV-TK was inserted between the two restriction sites of EcoRI and XhoI and its length was about 1200 bp. So theoretically, pHRE-Egr1-HSV-TK digested with EcoRI and XhoI can generate an about 1200 bp fragment of HSV-TK. Similarly, MluI and NheI digestion may lead to an about 1152 bp of Egr1 and BglII digestion can result in 380 bp fragment containing HRE. As shown in Fig  cells without transfection (Fig. 2B). These results suggested that HSV-TK was successfully integrated into the plasmids and transfected into HepG2 cells by PEI-MZF-NPs and can express stably at the mRNA level with radiation in hypoxic condition.

Effects of nuclide dose and radiation time of 131 I on gene expression.
For radiation-gene therapy study, ideally, radiation dose should be large enough to induce gene expression, but not be too high to damage target cells before the gene therapy takes effect. In addition, what time to heat is also important in the comprehensive treatment. In theory, as heat can inhibit the activity of some enzymes, impair DNA damage repair, and increase the permeability of cell membrane to take more drugs into cells, the effect should be better if heating is done firstly or thermotherapy is implemented with radiotherapy or chemotherapy simultaneously. But in the current study, it was radiation-gene therapy combined with thermotherapy that treated hepatocellular carcinoma, if so done, the cells would suffer from thermal damage from the very start, which would weaken the gene expression and consequently lower the effect of gene therapy. To explore the impacts of nuclide dosage and radiation time on gene expression, we checked the report gene EGFP expression in Bel-7402 cells induced by internal irradiation of 131 I at doses ranging from 50 to 400 μ Ci within 48 h in hypoxic conditions. The findings showed that EGFP expression had an obvious nuclide dose-effect and time-effect within 24 h and 200 μ Ci. 131 I irradiation with 200 μ Ci for 24 h made gene expression reach the peak. After 24 h or over 200 μ Ci, the expression started to decline (see Table 1). This indicated that large dose and (or) long irradiation injured the cells and thus weakened the gene expression. Therefore, the dose of 200 μ Ci might be the optimal dose for irradiation and the best time to initiate thermotherapy was 24 h irradiation later, which can ensure HSV-TK gene express sufficiently and make radionuclide, gene and hyperpyrexia yield a maximum comprehensive effect while each of them playing its own therapeutic role. This provides a theoretical evidence for choosing nuclide dose and when to start heat after irradiation in the later combination therapy.

Temperature rise of PEI-MZF-NPs/pHRE-Egr1-EGFP in vitro.
Undoubtedly, heat can improve the sensitivity of radiotherapy and chemotherapy, but the comprehensive curative effect is closely related to heating temperature. The critical temperature leading to death for most tumor cells is 42-43.5 °C, and heating alone at below 42 °C cannot effectively kill tumor cells, even overtime for heating is also of no useful. When temperature is more than 42 °C, thermal damage becomes obviously, and the quantity of killed cells can double when temperature rise by 1 °C 46,68 . But when it is more than 45 °C, the thermal sensitivity has no difference between tumor tissue and normal tissue, indicating damage on normal tissue. Therefore, the appropriate temperature for tumor thermotherapy should be over 42 °C, but not more than 45 °C, with an exception of good enough targetability.
In the present study, PEI-MZF-NPs/pHRE-Egr1-EGFP were dispersed in 0.9% NaCl and exposed to a high-frequency alternating electromagnetic field for 60 min. As shown in Fig. 2C, after exposure to the magnetic field, each fluid with different concentration could rapidly warm, and then became stable. As the concentration increased, the top temperature rose. Of them, 10 mg/l fluid rapidly warmed within 5 min, and then gradually reached a steady, the temperature approximately stabilizing at 43 °C or so, which is similar to magnetic induction temperature rise of single PEI-MZF-NPs or MZF-NPs tested in our previous study 16,17 , indicating that the combination with DNA does not change MZF-NPs' magnetic induction temperature rise characteristic. This hyperthermia behavior is quite appropriate for cancer treatment since it can kill tumor cells while not harming normal tissues. Thus, the dose of 10 mg/l was selected for magnetic induction hyperthermia in the later experiments.
HepG2 cell proliferation tested by MTT. Next, cell proliferation inhibition was explored. The cells of each group with different treatment were incubated over a period of 72 h and the cell viability was determined by MTT. As a classical chemotherapeutic widely used in clinic, adriamycin was chosen as a therapeutic control. As shown in Table 2, the combination greatly inhibited HepG2 cells' proliferation. The proliferation inhibition rate of cells treated with nuclide-gene-MFH is up to (94.51 ± 0.91)%, significantly higher than (43.91 ± 4.58)% of the nuclide-alone group, (60.33 ± 3.71) of the MFH-alone group, (74.97 ± 1.91) of the nuclide-gene group and (71.68 ± 2.06) of the adriamycin group (p < 0.001).

Flow cytometric analysis of apoptosis and necrosis. The HepG2 cells of each group after treatment
were stained with annexin-V-FLUOS/PI and then analyzed by flow cytometry to detect apoptosis and necrosis. The results showed the best apoptotic and necrotic efficacy in nuclide-gene-MFH group. The total apoptotis and necrosis rate of the nuclide-gene-MFH group reached 84.04%. In contrast, it was only 27.73% in the nuclide-gene group, 15.72% in the nuclide-alone group, 25.22% in the MFH-alone group, 57.03% in the adriamycin group, and only 6.55% in the blank control group (Fig. 3).
In vivo therapeutic effects of pHRE-Egr1-HSV-TK/ 131 I-antiAFPMcAb-GCV combined with MFH on xenograft hepatoma in nude mice. To investigate heating effect, antitumor efficacy and safety in vivo, hepatoma models were established in subcutaneous tissues of nude mice.
Heating in vivo. Following intratumor injection of PEI-MZF-NPs or PEI-MZF-NPs/pHRE-Egr1-HSV-TK/ 131 I-antiAFPMcAb-GCV and exposure to high frequency alternating magnetic field (AMF) for 30 min, the entire tumors almost got heated in MFH-alone group and nuclide-gene-MFH group, and then the tumor temperature maintained at 42-45 °C. All nontumor heating levels were safely below 41 °C during treatment (Supplemental Figure 2).
The tumor inhibitory effects. After six weeks treatment, all the tumors were removed from the nude mice and measured in volume and weight. Compared with saline control group, the tumors of all therapeutic   Table 3). Seen from in vivo tumor growth curves (Supplemental Figure 3), nuclide-gene-MFH group also showed the best therapeutic efficiency. The growth of tumors treated with nuclide-gene-MFH was greatly inhibited, obviously better than that treated with saline and the other therapies.
Histopathological findings, cell ultrastructural examination and immunohistochemistry (IHC) assays. To further assess the mechanism of therapeutic effects, histology of the excised tumor were analyzed via H&E staining and cell ultrastructure was detected by TEM. The H&E staining assays showed some magnetic nanoparticles (marked by the arrows) cumulated in the tumor tissue of MFH-alone group and nuclide-gene-MFH group. Widespread tumour necrosis with nucleus collapse and many inflammatory cells emerged in nuclide-gene-MFH group, much more than that in the other therapeutic groups, while tumor cells in saline control group were dense, with plenty cytoplasm and deep-dyed big nucleus (Fig. 4). These findings were consistent with the results of in vitro antitumor assays and in vivo tumor growth inhibitory effects.   As an active cellular suicide, apoptosis is often characterized as chromatin margination, nuclear fragmentation, cytoplasmic blebbing and internucleosomal fragmentation of DNA. Cell ultrastructure examination by TEM verified that nuclide-gene-MFH effectively induced xenograft hepatoma cells apoptosis. As shown in Fig. 5, hepatoma cells in nuclide-gene-MFH group (Fig. 5B,C) exhibited typical morphologically features of apoptotic cells such as karyopyknosis, chromatin condensation and margination or cleavage, apoptotic body, and cytoplasmic vacuolation (marked by the black arrows), and some magnetic materials (marked by the white arrow) deposited in the cytoplasm. By contrast, hepatoma cells in saline control group (Fig. 5A) were regular shape, with intact nucleus, fine chromatin and large nucleolus.
To further confirm the proliferation inhibition and apoptotic induction of the combination therapy, immunohistochemistry (IHC) assay were applied to detect Ki67 and survivin protein expression. As the strongest apoptosis inhibition factor ever discovered, survivin is closely related to the development and prognosis of cancer. It is barely expressed in normal tissues, but over-expressed in many tumors such as hepatoma, gastric cancer. The over-expressed survivin protein can inhibit apoptotic protease, help tumor cells escape from the checkpoint monitoring of cell cycle G2/M transition and resist cell apoptosis from DNA damage or mutation, resulting in abnormal cell division and proliferation 69 . Ki-67, a nuclear proliferation marker, is one of the most reliable indictors to assess the proliferative activity of tumor cells. It can be applied to evaluate the efficacy of cancer treatment. The more Ki-67 expresses, the brisker the cells grow 70,71 . Seen from the IHC assays ( Fig. 6 and Table 4   were both merely 1 which means extremely faint positive. This finding indicates that this combination may inhibit the tumor cells proliferation via inhibiting Ki67 protein expression and induce tumor cells apoptosis via inhibiting survivin protein expression. The above datas indicate that pHRE-Egr1-HSV-TK/ 131 I-antiAFPMcAb-GCV/MFH may make a good complemental synergetic anticancer effect, reaching 1 + 1 + 1 > 3. The mechanisms of action likely include the followings: 17 (1) Heat damages the repair of DNA fractured by radionuclide; (2) High temperature destroys the biological integrity of cell membrane and increases its permeability, which is helpful to the chemical drug and gene penetration and absorption; (3) G2 and M phase cells are very sensitive to radiation, whereas S phase cells are resistant to radiation but very sensitive to heat. Thus, nuclide radiotherapy and hyperpyrexia therapy can complement each other; (4) Heat rouses some dormant cells (G0 phase) into proliferation stage through improving the cell metabolism level. These dormant cells are usually resistant to radiation, but can become sensitive to radiation once getting into proliferation stage. As well, the sensitivity of original proliferative cells is further enhanced; (5) Heating can improve the radiation sensitivity of hypoxic tumor cells by improving the blood and oxygen supply to tumor region. Taken together, radionuclide, suicide gene and MFH play distinct roles in antitumor therapy. The three combination may overcome their own flaws and make their strengths complementary, leading to a synergistic effect. In addition, targeted treatment guided by antiAFPMcAb may also contribute to the good therapeutic effect.
In vivo safety evaluation. It is well known that chemotherapy and radiotherapy frequently result in side effects such as liver and kidney function damage, blood cells decrease caused by inhibiting bone marrow hematopoietic, and so on. For suicide gene therapy, non-toxic precursor drugs only can be converted to active substances in the place where suicide genes express, therefore, in theory, as long as the targeting of gene delivery is good enough, the treatment will not cause systemic side effects. In comparison, thermotherapy is considered a relatively safe and reliable therapeutic strategy which patients are nearly able to tolerate. Moreover, heat treatment can improve patient immunity and reduce adverse effects caused by chemotherapy and radiotherapy.
In the present study, the safety in vivo of the combination targeted therapy against hepatoma was preliminarily evaluated. We observed the survival of nude mice and tested AST, ALT, BUN and Cr levels in plasma and counted  peripheral blood cells of nude mice treated by pHRE-Egr1-HSV-TK/ 131 I-antiAFPMcAb-GCV/MFH, compared with those of normal nude mice without tumor, nude mice with tumor but without treatment and nude mice treated with the other therapies. In addition, we examined histological changes of heart, liver, spleen, lung, kidney, brain, and pancrease by H&E staining. We found no mouse death during treatment. AST, ALT, BUN and Cr levels ( Table 5) and the quantity of WBC, RBC, PLT (Table 6) did not change appreciably in any group, with no significant differences in the values among the seven groups (p < 0.05), suggesting no adverse effects on liver function, kidney function and no inhibition on marrow hematopoiesis of mice. The histological findings also exhibited no obvious pathological abnormalities in heart, liver, spleen, lung, kidney, brain, and pancrease tissues based on H&E staining (Fig. 7). These results indicated that the targeted combined therapy is safe and reliable. Of course, more safety indicators need to be further assessed. All this serves as a scientific theory and a sound basis upon which further tumor treatments and studies in clinic will be conducted in the future.

Conclusion
In this study, we successfully combined radionuclide, suicide gene and MFH organically to treat hepatoma targetly by using PEI-MZF-NPs as a linker. The in vitro and in vivo experimental results demonstrate that this combined hepatoma-targeted therapy has a good therapeutic efficacy, far better than any monotherapy and adriamycin chemotherapy. Furthermore, this combined therapy has few or no adverse effects. Thus, it offers a new effective and feasible strategy for hepatoma treatment.

To test the expression of HSV-TK by RT-PCR.
To determine the expression of HSV-TK, pHRE-Egr1-HSV-TK was transfected via PEI-MZF-NPs. In detail, PEI-MZF-NPs and pHRE-Egr1-HSV-TK were diluted in their separate serum-free medium (mass ratio of PEI-MZF-NPs/DNA was 40:1), then mixed together and incubated for 30 minutes at room temperature. Subsequently, pHRE-Egr1-HSV-TK/PEI-MZF-NPs complex was obtained. (2) HepG2 cells were seeded in 6-well plates at a density of 5 × 10 5 cells per well and incubated in routine conditions. About 18 hours later, the original culture medium was discarded, and the cells were rinsed twice with PBS and once with serum-free DMEM. (3) Serum-free DMEM with pHRE-Egr1-HSV-TK/ PEI-MZF-NPs was added to the wells (DNA: 3 μ g/well), and then the plates continued to be in the couveuse. 5 hours later, the serum-free medium was replaced by fresh serum DMEM medium, and then the cells continued to be incubated. The transfected cells were named HepG2/TK. After incubation for 24 h, the HepG2/TK cells were exposed to X-ray with 4 Gy (6 Mev) under a linear accelerator (SIEMENS, PRIMUS.HI), then continued to be incubated at 37 °C in hypoxic conditions (0.1% O 2 , 5% CO 2 and N 2 balance gas). 72 h later, the total RNA of the HepG2/TK cells was extracted. As well, the HepG2 cells without transfection were used as negative control, and GAPDH cDNA was used as internal reference. Primer sequences were synthesized by the Shanghai Ying Jun Company.  original culture medium was discarded, and the cells were washed twice with PBS and once with serum-free DMEM. (2) Serum-free DMEM with pHRE-Egr1-HSV-TK/PEI-MZF-NPs was added to the bottles (DNA: 3 μ g/bottle, PEI-MZF-NPs: pHRE-Egr1-EGFP = 40:1), and then the bottles were continued to be in the couveuse. 5 hours later, the serum-free medium was replaced with fresh serum DMEM medium, and then the cells were continued to be incubated.  (6) were cultured for 48 h in routine conditions. Subsequently, the cells in each bottle were digested with 0.25% trypsin and diluted into single cell with their corresponding original culture medium, then some of them were subcultured in 96-well plates (200 μ l/well), respectively. Group (3) and (5) were further incubated for 24 h under hypoxic conditions. Group (1), (2), (4) and (6) were ulteriorly cultured for 24 h in air 5% CO 2 at 37 °C. 24 h later, 20 μ l (5 g/l) of MTT was added to the cells in each well and continued to be incubated for 4 h. The culture medium was replaced with 150 μ l of DMSO and vibrated for 10 min. Then the optical density (OD) values were measured at a wavelength of 493 nm using a microplate reader (Multiskan MK3-353, USA). The cell proliferation inhibition ratio was calculated with the following formula: proliferation inhibition ratio (%) = (1 − OD of the experimental group/OD of the blank control group) × 100%. Three replicates were done in every group and each experiment was performed three times.
Flow cytometry assay. To evaluate the effect of the combination therapy on cell apoptosis and necrosis, the remaining cells of each group above (for MTT assay) were collected respectively for flow cytometry assay. After wash with PBS, supernatant in each group was decanted and each cell pellet was resuspended in 100 μ l staining solution which contained 2 μ l Annexin-V-FLUOS, 2 μ l propidium iodide (PI) and 96 μ l binding buffer, respectively. After 15 min incubated in dark at 25 °C, each cell sample was added with 400 μ l binding buffer and then analyzed by a flow cytometer (FCM, Vantage SE, BD Company, USA) within 1 h. Each experiment was repeated three times.
In vivo experiments of nude micebearing hepatocellular carcinoma. Female BALB/c nude mice, aged 6 weeks, weighing 20-22 gram, purchased from the Lakes Animal Experimental Center of the Institute of Biochemistry and Cell Biology, Shanghai Institute of Biological Sciences, China, were used for the experiments. The experiments were approved by the Animal Care Committee of Jiangsu Province and were performed in accordance with the institutional guidelines. All the mice were maintained in the sterile barrier system of Medical School, Southeast University, China. Exponentially growing HepG2 cells (2 × 10 6 cells) were injected subcutaneously around the right posterior limb rump.
Heating rise in vivo detection. During the mice of group (3) and (6) were heated on a high frequency heater coil plate, the temperature of tumor and nontumor region was measured using an infrared thermometer (ZyTemp-TN18 model, China) at 5 min intervals, respectively. Using the temperature as ordinate and time as abscissa, heating curves in vivo were drawn.
Tumor growth inhibition test. During treatment, the volumes of tumors were measured every 4-7 days, which were calculated as V = ab 2 /2 (The a is the long diameter and the b is short diameter). The tumor growth curves in vivo were drawn, using treat time as abscissa and tumor volume as ordinate. After treatment, the volume of tumors removed from mice were also measured and calculated as the above formula. In addition, the mass of tumors were weighted. Tumor growth inhibition ratio was evaluated by measuring mass and volume inhibition percentage. Mass inhibition ratio = (1 − relative tumor mass) × 100%, where relative tumor mass was the tumor mass of the experimental group divided by the mean tumor mass of the saline control group. Volume inhibition ratio = (1 − relative tumor volume) × 100%, where relative tumor volume was the tumor volume of the experimental group divided by the mean tumor volume of the saline control group.
Histology and cell ultrastructure examination. After mass and volume measurement, some removed tumors were fixed in 4% neutral formaldehyde, then sectioned for HE histopathological examination. Some removed tumors were fixed in 4% glutaraldehyde and then made into ultrathin sections for cell ultrastructure detection by transmission electron microscope (TEM) (JEM-200CX, Japan). Some other removed tumors were embedded in paraffin, then cut into sections (4 μ m in thick) for Ki67 and survivin IHC analysis.
To assess the safety, hearts, livers, spleens, lungs, kidneys, brains, and pancreases removed from mice of the radionuclide-gene-MFH group were fixed in 4% neutral formaldehyde and then sectioned for HE histological detection.

Blood cells and biochemical examination.
To evaluate impacts on marrow hematopoiesis function, white cells (WBC), red cells (RBC) and platelets (PLT) in the blood of mice in each group were counted by an automatic blood cell analyzer (PERLONG MEDICAL, XFA6100, China). After blood samples were centrifuged and separated, alanine aminotransferase (ALT), aspartate aminotransferase (AST), urea nitrogen (BUN) and creatinine (Cr) levels in the plasma were detected by a biochemical autoanalyzer (Beckman-LX20, USA) for liver and kidney function assessment.

Statistical analysis.
Values are shown as mean ± SD. The data were analyzed with the SPSS 16.0 program. A p value of < 0.05 was considered significant.