Platinum nanoparticles labelled with iodine-125 for combined “chemo-Auger electron” therapy of hepatocellular carcinoma

Convenient therapeutic protocols against hepatocellular carcinoma (HCC) exhibit low treatment effectiveness, especially in the context of long-term effects, which is primarily related to late diagnosis and high tumor heterogeneity. Current trends in medicine concern combined therapy to achieve new powerful tools against the most aggressive diseases. When designing modern, multimodal therapeutics, it is necessary to look for alternative routes of specific drug delivery to the cell, its selective (with respect to the tumor) activity and multidirectional action, enhancing the therapeutic effect. Targeting the physiology of the tumor makes it possible to take advantage of certain characteristic properties of the tumor that differentiate it from other cells. In the present paper we designed for the first time iodine-125 labeled platinum nanoparticles for combined “chemo-Auger electron” therapy of hepatocellular carcinoma. High selectivity achieved by targeting the tumor microenvironment of these cells was associated with effective radionuclide desorption in the presence of H2O2. The therapeutic effect was found to be correlated with cell damage at various molecular levels including DNA DSBs and was observed in a dose-dependent manner. A three-dimensional tumor spheroid revealed successful radioconjugate anticancer activity with a significant treatment response. A possible concept for clinical application after prior in vivo trials may be achieved via transarterial injection of micrometer range lipiodol emulsions with encapsulated 125I-NP. Ethiodized oil gives several advantages especially for HCC treatment; thus bearing in mind a suitable particle size for embolization, the obtained results highlight the exciting prospects for the development of PtNP-based combined therapy.


Introduction
It is generally accepted that metastasis is one of the leading reasons for recurrence and the consequent cancer mortality. 1 Dissemination of cancer sites consisting of single cells throughout the whole organism frequently located far away from the primary tumor site makes effective treatment by surgery or external radiotherapy impossible. Current therapeutic protocols include systemic chemotherapy as a major treatment option to deal with metastatic sites formed during cancer progression. The risk of severe side effects oen makes it difficult or unreasonable to continue the treatment, especially when a high tumor burden on critical organs occurs. Therefore, there is an urgency to investigate effective therapies that can be primarily oriented towards micrometastases.
Targeted radionuclide therapy (TRT) based on short-range radiation-emitting radionuclides improves the treatment of such disseminated cancers by using radiolabeled compounds that specically target tumor cells. These conditions are perfectly met by a and Auger emitters (AEs), high linear energy transfer (LET) radionuclides, which exhibit impressively precise and effective radio cytotoxicity, particularly by lethal DNA double-stranded breaks (DNA DSBs). 2 The tissue range of a-emitters and AEs is limited to no more than several cells (a, 40-100 mm) or even restricted to single cells (AE, 1-20 nm). Unfortunately, despite achieving spectacular therapeutic effects, 3 a-therapy cannot be used more widely due to the low availability of a-emitters. Current supplies of radioactive actinium ( 225 Ac, ∼1.7 Ci per year), which is the most popular a-emitter in nuclear medicine, are sufficient only for preclinical studies and for the therapies of a few hundred patients worldwide. Recent intensive work on the cyclotron production of 225 Ac in the spallation reaction on thorium ( 232 Th(p,x) 227,225 Ac) and by the proton irradiation of radium (Ra; 226 Ra(p,2n) 225 Ac) did not produce the expected results, mainly due to the contamination of 225 Ac by the long half-life of 227 Ac (t 1/2 = 20 years). 4,5 AEs may be a noteworthy alternative to achieve precise cell targeting in TRT. Improved availability and radiochemical properties can make AEs an excellent tool for enhancing the selectivity of treatment on the molecular level of an individual cell. The nuclear pathway of AEs involves decay via electron capture (EC) or internal conversion (IC) which leads to the formation of vacancies in the electronic shells. Rapid lling by electrons relaxing from external orbitals is accompanied by characteristic X-rays or cascades of Auger electrons. Moreover, these characteristic X-rays and Auger electrons offer an encouraging prospect for inducing a signicant therapeutic response. Contrary to a-emitters, there is a wide range of Auger electron emitters that can be applied in TRT. The most commonly used AEs are radioactive indium (In; 111 In), iodine (I; 125 I), gallium (Ga; 67 Ga), and technetium (Tc; 99m Tc). Because all of these radionuclides emit g-quanta in high yield, they are used extensively in radiodiagnosis. They are also commonly available and well-characterized.
Auger electron emitters located in the DNA were found to be more radiotoxic than the a particle emitter polonium (Po; 210 Po). 6 Furthermore, in contrast to a and b − radiation, AEs remain nontoxic while traveling in the blood or bone marrow. Conversely, AEs become highly efficient when located in the cell nucleus and incorporated into the DNA of the target cells. For these reasons, Auger radiotherapy is considered a promising emerging eld in nuclear medicine, especially for targeting individual cancer cells and small tumor metastases as well as post-operative supportive therapy. 7,8 The main restriction for the clinical application of AEs is the design of appropriate delivery systems which meet the rigorous criteria for the efficient and selective delivery of the high activities of Auger-based radiopharmaceuticals. Because most of the energy released by Auger electrons is deposited within proximity to the decay site, effective intracellular uptake is a minimum requirement, while targeting more sensitive organelles (e.g. mitochondria or the cell membrane 9 ) is preferable. However, the most vulnerable target for an AE is DNA. Thus, nuclear uptake is another prerequisite for improving therapeutic efficacy. Apart from the direct interaction of Auger electrons with DNA, single-and double-stranded DNA breaks can also be indirectly induced by hydroxyl radicals (cOH) and other reactive oxygen species (ROS) generated by Auger electrons in the cytosol. Therefore, multifunctionality is another advantage of this concept.
The therapeutic effect is strongly correlated with the high specic activity of a designed radiopharmaceutical. Direct attachment of radionuclides to biomolecules through chelators does not result in sufficient specic activities. For example, the specic activity of radiolabeled antibodies was relatively small (0.24 MBq mg −1 ), but several different strategies have been implemented to overcome this shortcoming. 10,11 Moreover, many scientic reports have investigated the use of gold (Au) nanoparticles (NPs; AuNPs) as carriers of AEs. [12][13][14][15][16] In our work, we propose to use platinum nanoparticles (PtNPs) and platinum coated gold core-shell nanoparticles (Au@Pt NPs) as transporters of Auger radiation emitters. Metastable Pt radionuclides ( 193m Pt and 195m Pt) are one of the most effective Auger electron emitters with great potency for Fig. 1 Postulated mechanism of combined "chemo-Auger electron" therapy. medical applications. Unfortunately, it is almost impossible to obtain sufficient activities to carry out the therapy. 17,18 According to the most promising procedure, 19 we investigated the possibility of producing this nuclide indirectly, via neutron irradiation using an iridium target according to the two step reaction: 193 Irðn;gÞ 194 Irðn;gÞ 195m Ir ! b À 195m Pt. Weekly irradiation of the 193 Ir target in the Maria reactor inŚwierk (Poland) with a 1.5 × 10 14 n cm 2 s −1 neutron ux yielded only ∼100 MBq for 1 mg of the target. Due to side reactions resulting in 192 Ir and 194 Ir production and difficulties in processing the target material, increasing the target mass and thus implementing this method was not reported.
Therefore, in our work, we did not pursue the application of 195m Pt. Instead, another effective AE, 125 I, was utilized. For 125 I decay associated with DNA, a "one decay = one DNA DSB" rule was postulated, 20,21 which should correlate with high therapeutic effectiveness. To attach 125 I to the surface of NPs we utilized the strong affinity of iodine atoms for adsorption on the surface of noble metals. We used 125 I-coated PtNPs/Au@Pt NPs for hepatocellular carcinoma (HCC), the most common type of primary liver cancer in adults and currently the third leading cause of cancer-related deaths worldwide. 22 We chose this type of cancer because HCC cells (HepG2) contain abnormally high levels of hydrogen peroxide (H 2 O 2 ) 23, 24 and we hypothesized that in the presence of H 2 O 2 , 125 I oxidation will occur on the surface of Au@Pt NP/PtNP, and 125 I will easily release from the NP and freely diffuse near the DNA. Therefore, the proposed system should meet the criteria that are necessary for effective Auger electron therapy. In addition, in our previous work, 25 we noted the high toxicity of nonradioactive Au@Pt NPs/PtNPs towards these cells related to the catalytic decomposition of H 2 O 2 into more reactive ROS radicals on the NP surface and intranuclear uptake of PtNPs. Thus, such a system should enable selective chemo-and radiotoxic effects on HepG2 cells without a targeting moiety (Fig. 1).
For biological experiments the following materials were used: EMEM; trypsin EDTA solution C; water, cell culture grade; phosphate-buffered saline (PBS), and fetal calf serum from Biological Industries (Beth Haemek, Israel). The CellTiter96® AQueous One Solution Reagent (MTS compound) from Promega (Mannheim, Germany). HepG2 cells were obtained from the American Type Tissue Culture Collection (ATCC, Rockville, MD, USA) and cultured according to the ATCC protocol. For experimental applications, over 80% conuent cells were used. All solutions were prepared with ultrapure deionized water (18.2 MU cm, Hydrolab, Straszyn, Poland).

Chemisorption of iodine radionuclides and stability studies
During each synthesis, the following amounts of nanoparticles (synthesized and characterized as we previously described 25 ) were applied: 4.5 × 10 10 (AuNPs), 1.9 × 10 10 (Au@Pt), and 3.95 × 10 14 (PtNPs) to have a similar outer surface area in every sample. First, for preconcentration and purication, nanoparticles were centrifuged and redispersed in deionized water. For 30 nm gold and core-shell nanoparticles, centrifuging for 10 min and 10 000 rpm was implemented, while for platinum nanoparticles, centrifugal concentrators Vivaspin®500 (cut-off MW = 3 kDa) were used for 20 min and 13 400 rpm. Subsequently, the radionuclide was added and the reaction was kept at room temperature for 1 hour with continuous mixing. Finally, HS-PEG-COOH (MW = 5 kDa) was added with the desired excess and the reaction was continued for the next 30 min.
Quality control was performed with the instant thin layer chromatography (iTLC) technique with the use of Storage Phosphor System Cyclone Plus (PerkinElmer, Waltham, MA, USA), glass microber chromatography paper impregnated with silica gel (iTLC SG, Agilent Technologies, Santa Clara, CA, USA), and methyl alcohol (MeOH) as the mobile phase. Analyses were performed aer the rst and nal steps of radioconjugation as well as, during stability studies. For additional radiochemical yield evaluation, we centrifuged nanoparticles aer each synthesis step and measured the activity of the collected fractions.
For stability studies radioconjugates were puried aer synthesis and redispersed in 100 mL of deionized water. Next, 300 mL of human serum (HS) or phosphate-buffered saline (PBS) were added and the samples were heated at 37°C for three days.
Iodine release in a highly oxidative environment Selected concentrations of hydrogen peroxide were prepared by dissolution of stock solution (30%) with deionized water. Then, radioconjugates aer prior centrifugation, were dispersed directly in prepared H 2 O 2 solutions. Samples were then incubated for 1 h at 37°C. Aer this, quality control with iTLC was performed to nd the released iodine ratio, as described above.

In vitro toxicity
HepG2 and HeLa cells (8 × 10 3 ) were seeded into 96-well plates and incubated overnight (37°C, 5% CO 2 ). Next, the growing medium was removed and radioconjugates in 100 mL of fresh medium were added. Aer 24, 48, and 72 h incubation, the medium was removed and a fresh one was added. Finally, 20 mL of CellTiter96® AQueous One Solution Reagent (Promega, MDN, USA) was added and absorbance was measured at 490 nm to calculate the % of metabolically active cells.

DNA double-strand breaks
HepG2 cells (2.5 × 10 5 per well) were seeded into six-well plates with sterile glass coverslips (diameter 12 mm), Thermo Fischer Scientic (Waltham, MA, USA) and incubated overnight. Aer removing the medium, cells were treated with radioconjugates (0-100 MBq mL −1 ), iodine-125 (100 MBq mL −1 ), and staurosporine (0.5 mM) and incubated for 12 and 24 h. Furthermore, the protocol was analogous to that reported previously. 30 For gH2A.X foci detection with primary anti-phospho-Histone H2A.X (Ser139) antibody, clone JBW301 was dissolved 1 : 100 with blocking buffer (BB -4% BSA in TBS) and 350 mL per well was added for overnight incubation at 4°C. The next day, primary antibody was replaced with anti-mouse IgG secondary antibody conjugated with CF™ 633. Antibody was dissolved in blocking buffer according to manufacturer's protocol and the cells were incubated for 2 h at room temperature with mixing. Finally, the cells were washed 3 times with water and for nuclei staining, Hoechst 33 258 was used. Imaging was performed on an FV-1000 confocal microscope (Olympus Corporation, Tokyo, Japan) with ex/em maxima: 630/650 nm for CF633 and ex/em maxima: 352/454 nm for Hoechst 33 258. The results were analyzed with Fiji 2.9.0 version.

I-PtNP effects on a 3D tumor spheroid model
HepG2 cells (1 × 10 3 ) suspended in 200 mL of growing medium were seeded into 96-well U-bottom ultra-low adherent plates (Corning®, Corning, NY, USA) seven days before the experiment, as reported previously. 31 During incubation, 100 mL of medium was replaced every two days. Aer 7 days, radioconjugates in 100 mL growing medium were added. Spheroid images with area measurements were obtained and analyzed with a ZEISS Axio Vert.A1 Microscope and ZEN 2.1 soware (Zeiss, Jena, Germany).

Results and discussion
Chemisorption of 131 I on the nanoparticle surface Au and Pt are noble metals that exhibit similar atomic properties, such as covalent radius and electronegativity, which determine the type and strength of chemical bonds. A common feature of these metals is the formation of strong bonds with heavy halogens such as iodine and astatine (At). [32][33][34] The Pt-I bond length (261 pm) is slightly shorter than that of Au-I (277 pm), which suggests the formation of a stronger Pt-I bond than Au-I. Analogously to Au, the chemisorption reaction of I − onto the surface of Au@Pt NPs/PtNPs should be as follows (eqn (1)): In our work, we utilized the high affinity of Pt atoms for iodine atoms to immobilize 125 I on the surface of the NPs. Due to wider availability and cost-effectiveness, some of the chemical studies were performed on the 131 I radionuclide having the same chemical properties. The studies were performed on 2 nm PtNPs, Pt-coated 30 nm AuNPs (Au@Pt NPs) and 30 nm AuNPs for comparison. Post synthesis, 131 I was successfully immobilized on the NP surface with greater than 85% yield of chemisorption. Additionally, negligible changes occurred aer subsequent surface functionalization with polyethylene glycol (PEG; PEG-ylation) for NP stabilization, dispersity enhancement, and purication realized in the nal step (Table 1). As calculated, the average outer surface of a single NP accessible to iodine atoms was 12.5 nm 2 for a 2 nm PtNP and 2.8 × 10 3 nm 2 for a 30 nm Au@Pt NP. Considering the theoretical cross-sectional area of a single I atom, ∼18 I atoms could be deposited per 1 nm 2 of these NP surfaces. In our procedure, ∼0.9 iodine atoms per 1 nm 2 were used to avoid oversaturating the surface.
Interestingly, both Pt-containing NPs (Au@Pt NPs and PtNPs) exhibited decreased chemisorption yields when compared to AuNPs used as a control in a parallel experiment. This phenomenon can be explained by the different types of stabilizing agents directly conjugated to the NP surface. Each type of NP is stabilized using a diverse agent, such as sodium citrate (e.g., AuNPs), ascorbic acid (e.g., Au@Pt NPs), and sodium rhodizonate (e.g., PtNPs). Iodine chemisorption may occur aer I − oxidizes to I 0 (from Na 131 I aq ) and the subsequent direct deposition on the metal surface (eqn (1)), either via the replacement of stabilizing molecules 35 or a combination both of the mentioned alternatives. Therefore, it is reasonable to consider the type of stabilizing agent as an indirect cause of the diversi-ed yield for the different types of NPs investigated. Aer the PEG-ylation process, the release of 131 I by the competitive displacement of iodine atoms by the thiol group did not occur. Inconsistencies were limited to the statistical differences in each of the described radioconjugates.

Surface saturation
Assessing surface saturation was essential regarding the future application of 125 I-NP radioconjugates for therapeutic procedures. Since very high radionuclide activities are necessary for effective Auger electron therapy, a high specic activity is required to further consider 125 I-NP as a precursor for radiopharmaceuticals. 36 Incubation of NPs with increasing activity of 131 I indicated the formation of bound-to-free 131 I equilibrium in solution regardless of the concentration (Fig. 2).
Sorption on the AuNP surface was the most efficient (>90%) for various concentrations, up to 100 MBq of 131 I per 1 mL of NP solution. Au@Pt NPs and PtNPs exhibited inferior chemisorption efficiency. However, as observed, this is not related to the saturation and subsequent depletion of the NP surface available for the iodine. We observed the formation of an equilibrium state at each of the veried activities (20-100 MBq mL −1 ). Notably, the concentration of NP-bound I increased when the activity of 131 I was greater while the yield remained at a comparable level. This result could partially be due to the aermath of sideways reactions between the radionuclide and stabilizing agent, especially in the case of Au@Pt NPs and ascorbic acid, as reported previously. 37

Stability studies
The synthesized radioconjugates demonstrated different stabilities in phosphate-buffered saline (PBS) and human serum (HS) (Fig. 3). Almost complete iodine desorption (>90%) from PEG-ylated AuNPs occurred aer a longer than 1 h stability test. Pt-containing radioconjugates (Au@Pt NPs and PtNPs) revealed improved stability in HS when compared to AuNPs and reached ∼70% aer 48 h. PBS did not induce a loss of stability in 131 I-AuNPs and 131 I-Au@Pt NPs, whereas 131 I-PtNP stability remained at a constant level (75%) from 24 to 72 h. A lack of stability ( 131 I-AuNPs) or minorly reduced stability (e.g., 131 I-Au@Pt and 131 I-PtNPs) in HS can be related to the sulfurcontaining peptides and proteins present in the uid which can displace iodine from the NP surface. Like previously reported studies, 131 I-AuNPs exhibited decreasing stability over time in fetal bovine serum (FBS). 38 The stability of iodinated Ptcontaining NPs has not been previously studied. The much greater stability of iodinated PtNPs and Au@Pt NPs in HS conrms the prediction that the Pt-I bond is stronger than Au-I. All of the tested PEG-ylated and iodinated NPs did not agglomerate in both PBS and HS during long time point experiments. This is in line with the results of previous publications concerned with AuNPs. 39,40 Iodine release in a highly oxidative environment Due to their very short tissue range, AEs should be delivered into the cell nucleus, preferably close to the DNA. 41 As  previously reported, 25,42 2 nm NPs can penetrate the nuclear envelope of HepG2 cells making them very promising candidates for Auger electron therapy, especially when combined therapy can be considered. Successful delivery of iodine radionuclides closely or into the nucleus is also potentially achievable via iodine desorption in the cytosol of cancer cells with high oxidative status. Iodine reduction by H 2 O 2 is widely described in the literature. 43,44 In vitro this process leads to the direct formation of HOI 45 and other species via unstable intermediates. 46 For this purpose, we decided to study whether 125,131 I can be efficiently released from the NP surface under highly oxidative conditions corresponding to hepatic cancer cells (Fig. 4). Our previous studies concerned with designing novel HCC therapeutic approaches revealed that HepG2 cells have abnormally high oxidative status and signicantly reduced antioxidant protection capacity. 25 The physiological concentration of H 2 O 2 in various cells is in the nanomolar range; 47 thus 10 nM was considered a benchmark in this experiment. In order to verify whether iodine can be released in a dose-dependent manner in the presence of H 2 O 2 , we also implemented a higher concentrations range. Since we noticed that 125 I-AuNPs is not a suitable model for in vitro-related studies due to its poor stability in biological media (Fig. 3), we avoided applying this radioconjugate in our further studies as a reference compound.
In each of the investigated H 2 O 2 concentrations, iodine was effectively desorbed from PtNP radioconjugates reaching around ∼20% at 10 nM. The percentage of desorbed radionuclides increased with an increase in the H 2 O 2 concentration, highlighting the impact of the oxidizing agent in this process. Contrary to PtNPs, a negligible release was observed for Au@Pt NPs (∼5%), and any desorption was found to occur when radioconjugates were incubated in water. The signicant difference in iodine release between Au@Pt NPs and 2 nm PtNPs may refer to the partial dissolution of 2 nm PtNP in HepG2 cells postulated by Shoshan et al. 42 Another possible explanation for the observed differences is related to the diverse physicochemical properties of both types of nanoparticles. 2 nm PtNPs due to a higher SA : V ratio should be more reactive against H 2 O 2 as a result of their superior surface atom number compared to Au@Pt NPs. Both of the mentioned reasons (dissolution or reactivity) directly clarify iodine release occurring only in the case of 2 nm nanoparticles. If these circumstances occur aer internalization, then a noteworthy improvement in therapeutic efficacy could be achieved due to the signicant nuclear access for 125 I.
Considering the discussed treatment strategy with 125 I its major advantage in relation to 5-iodo-2 ′ -deoxyuridine labelled with 125 I ( 125 IUdR) must be emphasized. This most frequently occurring way to implement the therapy with 125 I without a targeting moiety has numerous side effects related mostly to systemic toxicity as a consequence of poor selectivity. Keeping in mind the necessity of overcoming this limitation, our strategy can be greatly implemented in future therapeutic approaches, as it can potentially ensure and improve cancer treatment selectivity.

In vitro toxicity
Investigation of radioconjugate-mediated cell viability effects was performed using HepG2 cells, while adenocarcinoma cells (HeLa) which do not exhibit oxidative properties were implemented as a control in this experiment. As we previously reported, oxidative status plays a key role in Pt-containing NP induced toxicity. Previously we observed that high oxidative potential is required for triggering Au@Pt NP/PtNP biological activity, as only HepG2 cells were affected by the treatment with various concentrations of NPs; cells with increased but moderate oxidative status cells (SKOV-3) remained unaffected. 25 Iodine release from PtNPs occurring only under highly oxidative conditions was another criterion taken into account in the selection of the in vitro model. Thus, following the strategy of targeting the tumor microenvironment, we veried the therapeutic potential of the synthesized radioconjugates against hepatic cancer cells (HepG2). HeLa cells, notwithstanding being a cancer cell line, are one cell line that is very similar to healthy cells. 48 Hence, we also evaluated the synthesized radioconjugates with HeLa cells as a reference cell line.
HepG2 cell viability (Fig. 5A) was not affected aer treatment with free 125 I − at any of the tested radioactivity concentrations (6.25-100 MBq mL −1 ). We also did not nd any radioactivityrelated response during the incubation with 125 I − Au@Pt NPs. The decrease in the mitochondrial activity fraction over time was related to the chemotoxicity of Au@Pt NPs with any enhancement of the effect achieved by electrons emitted from 125 I decay. This is a consequence of 125 I location outside of the nucleus, due to the intranuclear absence of Au@Pt NPs 25 and lack of I release (Fig. 4). Because of the characteristics of 125 I decay, therapeutic effectiveness was probably impossible to achieve at sites distant from the DNA. Decay of 125 I results in the emission of ∼23 AEs with an average energy in the range 0.07-30.1 keV and around 7.3 IC electrons with 35 keV maximum energy. 49,50 Comparing this with other Auger electron emitters, it was found that the low energies of the emitted particles make them unfavorable for expecting a high biological response induced by nuclides located away from DNA. 51 A signicant decrease in the mitochondrial redox activity was observed at 50 MBq mL −1 and 100 MBq mL −1 aer 24 h of incubation with 2 nm NPs (p # 0.01). Bare PtNPs (0 MBq mL −1 ) induced a ∼50% decrease in cell viability while 125 I-PtNPs improved the therapeutic effect to ∼90%. It has also been observed that 25 MBq mL −1 exhibits a slightly additive effect, but contrary to 50 and 100 MBq mL −1 , the effect was much weaker with a ∼10% improvement in the cell viability decrease when compared to PtNPs without radioactive I (p > 0.05). These effects were triggered by prooxidative activity of bare PtNPs and radiotoxicity of 125 I. Auger and IC electrons could induce signicant treatment enhancement through the intranuclear uptake of PtNPs of which ∼13-15% are located in the HepG2 cell nucleus aer 24 h 25 or due to iodine release from the nanoparticle surface. Core-shell nanoparticles' presence in the nuclei fraction was not detected reaching only ∼2% and oscillating within the error limit. These data directly demonstrate that AEs are most effective when localized in direct proximity to DNA. 52 It is applicable, especially when weaker AEs are considered. 53 HeLa cells were not affected by Au@Pt NPs or PtNPs (Fig. 5B). Application of 0-100 MBq mL −1 125 I did not decrease the mitochondrial activity of HeLa cells. In contrast to HepG2 cells, the low concentration of H 2 O 2 and resulting from other anatomical functions of primary tissue a different oxidative potential in HeLa cells did not cause the platinum biological activity and release of 125 I with subsequent penetration into the nucleus or DNA. The presented ndings support the hypothesis that the designed system is highly specic since its biological activity is triggered only under favorable conditions in cytosol.
Moreover, in agreement with our predictions, short-range AEs, even considering high LET and high activity doses, are unable to kill the cells without the precise targeting DNA achieved using a targeting vector or cell sensitivity expressed under specic physiological conditions. Cytotoxicity studies revealed that sorption of iodine does not impair biological activity of PtNPs, maintaining their pro-oxidant capacity. The additive effect observed in HCC cells from oxidative stress induced with PtNPs and radiotoxicity from 125 I gives an exciting opportunity to implement this strategy as combined therapy. Translation of these results into in vivo and preclinical evaluation needs to answer the question about the safety tolerance limit of nanoparticle concentrations. The range (145 mg mL −1 ) used in our current and previously reported studies is relatively low when compared to the data published by other groups. 54,55 Benecially, our concept to implement 125 I-PtNPs gives an excellent opportunity to obtain high specic activities required during the further development of this strategy in order to achieve satisfactory parameters for nuclear medicine.

Morphological changes
Treatment with radioconjugates changed not only cell mitochondrial activity but also affected alterations in cellular morphology. Microscopic images of cells incubated for 24 h with various concentrations of 125 I-PtNPs revealed substantial morphological changes in HepG2 cells when compared to nonradioactive PtNPs, 125 I (100 MBq mL −1 ), and untreated cells (Fig. 6). Increasing concentrations of 125 I-radioconjugates have resulted in intensifying modications of cellular morphology. HepG2 is a well-known cell line that typically grows by forming clusters and layers of cells. Treatment with radioconjugates disrupted (25 MBq mL −1 ) or destroyed (50 and 100 MBq mL −1 ) the integrity of these clusters. Increasing concentrations also prompted cell membrane shrinkage and a distinctive reduction in the cell number. Small changes were also observed in the control cells incubated with non-radioactive PtNPs, consisting of a minor reduction in the size of clusters followed by the separation of single cells outside from groups. This was a consequence of the chemotoxicity of PtNPs achieved by oxidative stress and intranuclear location, which was previously reported. 25 As expected, no modications were found in the cells treated with 100 MBq mL −1 of 125 I according to the MTS assay (Fig. 6). The low range of AE and IC electrons makes Auger therapy poorly or not effective without nuclear location, even for high LET particles, such as Auger electrons.

DNA double-strand breaks (DNA DSBs)
Despite progressive dysfunction of the mitochondria followed by signicant morphological changes, induction of DNA DSBs is considered one of the most desired outcomes of radiopharmaceutical anticancer activity. 56 Lethal and unrepairable damage of the genetic material is the most effective way to achieve a high therapeutic response. A growing number of DSBs in HepG2 cells was found with an increase in the radioactivity concentration, mostly in the range of 12.5-100 MBq mL −1 (Fig. 7A).
The average counts of phosphorylated H2A.X Histone in a single nucleus were 14 and 26 at 100 MBq mL −1 aer 12 and 24 h, respectively. We also observed up to 5 spontaneous DSBs in untreated cells. Aer 12 h of treatment, only a slight increase in gH2A.X foci was observed for most of the analyzed radioconjugate concentrations, when compared to 125 I (100 MBq mL −1 ). Longer incubations (24 h) revealed an increasing signal of DNA damage in a dose-dependent manner, primarily for 25 MBq mL −1 , 50 MBq mL −1 , and 100 MBq mL −1 (Fig. 7A and B) which corresponds to the data obtained during cytotoxicity investigation. Considering the redox activity of PtNPs, we could also expect an increased DSB ratio in cells treated with the nonradioactive compound. It is widely reported that ROS may induce serious DNA damage, including DNA DSBs. 57,58 In our studies, the level of gH2A.X foci in the cells treated with PtNPs was similar to that in untreated control cells. This can be explained by the kinetics of oxidative stress induced by applied NPs. First, low ROS overproduction was observed aer 12 h with a subsequent large increase from 24 to 72 h. 25 Thus, we conclude that 24 h of incubation was too short of a treatment for efficient DNA DSB generation via free radicals produced by nonradioactive PtNPs in the cytosol. The radioconjugate uniformly affected the treated cells, and the red signal from gH2A.X was homogenously detected in most of the cells (Fig. 7B and C). This is a very valuable outcome, as it conrms that the observed desirable signals are not accidental and not locally situated. Our results agree with those of previously reported models and calculations related to the ability of AEs to induce DNA DSBs. 59,60 However, the radionuclide-DNA distance plays a crucial role in effective DNA damage induction, which explains the lack of an effect from radioconjugates located outside of the nucleus, such as 125 I and 125 I-Au@Pt NPs. 61 Fig. 6 Morphological changes in HepG2 cells induced by nonradioactive (0 MBq mL −1 ) PtNPs as a "chemotoxic effect" and radioconjugates (25-100 MBq mL −1 ) as presentation of a dose-dependent manner of the induced chemo-Auger electron effect. Taking into account the low tissue range of AEs and the uncertain penetration of tumors by NPs, AEs, and IC electrons, we examined whether the synthesized radioconjugates would be effective against a 3D spheroid model. As we previously reported, Au@Pt NPs and PtNPs can effectively induce spheroid toxicity, but any decrease in the tumor area was related to high chemotoxicity revealed with uorescence imaging. 25 Since the expected therapeutic effect improvement in the case of 125 I-Au@Pt NPs was not achieved, we focused our attention on 125 I-PtNPs.
Because of 125 I characteristics and cytotoxicity studies (Fig. 5), we applied 50 and 100 MBq mL −1 of 125 I-PtNPs for continuous treatment until signicant tumor area reduction was distinguished. Fig. 8 describes spheroid area changes in treated and control samples. We observed that the rst symptoms of tumor growth inhibition were induced 72 h post-NP injection. These changes progressed more slowly when compared to the MTS assay, which was reasonable and expected. The spheroid area was then reduced by ∼10 and 15% for 50 MBq mL −1 and 100 MBq mL −1 , respectively, in reference to control cells. However, during this time, tumors maintained their initial structure without any morphological uctuations. The rst meaningful changes in the tumor shape and integrity were noted on the 5th day. In both of the tested concentrations, many cells detached from the tumor and ∼50% of the area decreased. Subsequently, 100 MBq mL −1 of radioconjugate led to the total disintegration of the spheroid on the 12th day, whereas 50 MBq mL −1 reduced tumor size by approximately sevenfold.

Conclusions
Due to the insidious growth nature of HCC, most patients are diagnosed at advanced stages of the disease, at which point the available therapeutic options are limited and ineffective.
Cutting-edge research and development in HCC nanomedicine has provided a powerful tool compared to traditional methods such as systemically applied sorafenib, transarterial chemo, and radioembolization. The combination chemo-Auger therapy showed additive chemotoxic effects of PtNPs and a radiotherapeutic action of Auger electrons emitted from 125 I occurring only in an elevated oxidative environment of HCC. Release of 125 I from the surface of PtNPs as a result of oxidation promotes better accessibility of the radionuclide to the structures critical for maintaining cell functions (e.g. the cell nucleus and DNA). Therefore, a selective effect of 125 I-PtNPs on HepG2 liver cancer cells is observed.
Prospective application may be realized by the transarterial injection of 125 I-NP lipiodol emulsions which have a suitable particle size for the embolization of the middle and small arteries of HCC. In addition, lipiodol has the advantage of selectively accumulating in HCC tissues, which enables 125 I-NP lipiodol emulsions to be efficiently absorbed by tumor tissues. What we propose is to apply 125 I-PtNPs as a drug loaded into micrometer range lipiodol emulsions allowing embolization with further chemo-Auger induced toxicity. Our results highlight the exciting prospects for the development of Pt NP-based therapeutics as highly effective, safe for healthy tissues and selective tools. Versality of the proposed radioconjugates gives an opportunity to personalize the treatment strategy through the selection of the administration route and may be associated with using SPECT properties of 125 I to image not only the radioconjugate location, but also the following treatment progress.

Conflicts of interest
There are no conicts to declare.