Bone tumor–targeted delivery of theranostic 195mPt-bisphosphonate complexes promotes killing of metastatic tumor cells

Platinum-based drugs such as cisplatin are very potent chemotherapeutics, whereas radioactive platinum (195mPt) is a rich source of low-energy Auger electrons, which kills tumor cells by damaging DNA. Auger electrons damage cells over a very short range. Consequently, 195mPt-based radiopharmaceuticals should be targeted toward tumors to maximize radiotherapeutic efficacy and minimize Pt-based systemic toxicity. Herein, we show that systemically administered radioactive bisphosphonate-functionalized platinum (195mPt-BP) complexes specifically accumulate in intratibial bone metastatic lesions in mice. The 195mPt-BP complexes accumulate 7.3-fold more effectively in bone 7 days after systemic delivery compared to 195mPt-cisplatin lacking bone-targeting bisphosphonate ligands. Therapeutically, 195mPt-BP treatment causes 4.5-fold more γ-H2AX formation, a biomarker for DNA damage in metastatic tumor cells compared to 195mPt-cisplatin. We show that systemically administered 195mPt-BP is radiotherapeutically active, as evidenced by an 11-fold increased DNA damage in metastatic tumor cells compared to non-radioactive Pt-BP controls. Moreover, apoptosis in metastatic tumor cells is enhanced more than 3.4-fold upon systemic administration of 195mPt-BP vs. radioactive 195mPt-cisplatin or non-radioactive Pt-BP controls. These results provide the first preclinical evidence for specific accumulation and strong radiotherapeutic activity of 195mPt-BP in bone metastatic lesions, which offers new avenues of research on radiotherapeutic killing of tumor cells in bone metastases by Auger electrons.


Introduction
Auger electrons are low-energy electrons emitted by specific radionuclides, which decay by electron capture. Auger-emitting radiopharmaceuticals are promising candidates for cancer treatment because the energy is deposited over a very short range in the order of nanometers [1]. 195m Pt is known for its high Auger emission intensity since this radionuclide emits 36 Auger electrons per decay, which exceeds the number of Auger electrons emitted by 111In (7Â), 123I (14Â), and 125I (23Â) by far. Consequently,195m Pt-based Auger therapy would enable potent effective killing of tumor cells if delivered within the vicinity of a tumor cell, as the energy deposited per decay is much higher for 195m Pt (2000 eV) as compared with 111 In (450 eV), 123 I (550 eV), or 125 I (1000 eV) [2].
Platinum (Pt)-based drugs are widely used for the treatment of 24 specific types of cancer [3]. Traditional Pt drugs such as cisplatin consist of two non-leaving amine groups and two additional leaving ligands that can bind DNA to induce DNA damage [4]. Cisplatin comprising radioactive Pt core ( 191 Pt,193m Pt, and 195m Pt) has been explored between 1990 and 2000 as an alternative to cisplatin for both tumor imaging and therapy [5,6], since cisplatin also acts as a radiosensitizer by enhancing the efficacy of radiation therapy [7]. However, clinical translation of radioactive Pt was hampered by its non-targeted tissue uptake leading to severe side-effects such as nephrotoxicity.
Fortunately, a novel generation of bone tumor-targeting Pt-based drugs has been developed recently to minimize toxicity (e.g. nephrotoxicity) and maximize therapeutic antitumor efficacy [8][9][10]. Margiotta et al. functionalized Pt with bisphosphonate ligands (Pt-BP) to facilitate the design of Pt-based chemotherapeutics targeting bone of high metabolic activity such as bone metastases [11][12][13][14]. These Pt-BP complexes were shown to act as prodrugs by dissociating into free bisphosphonate ligands and chemotherapeutically active Pt-species able to kill cancer cells in a manner similar to BP-free Pt drugs following their dissociation (Fig. 1A) [14].
Previously, we successfully synthesized radioactive 195m Pt-BP complex to facilitate targeting of Auger-emitting 195m Pt to bone metastases [15]. This novel radioactive complex accumulated specifically in bone of high metabolic activity. For optimal treatment of bone tumors, these 195m Pt-BP complexes should release 195m Pt species specifically at the tumor site to enable uptake by tumor cells, reaction with DNA, and formation of 195m Pt-DNA adducts to trigger tumor cell killing. Release kinetics of Pt species from Pt-BP complexes attached to hydroxyapatite mineral are typically slow, but are accelerated at reduced pH [12,14,15]. Therefore, we hypothesize that targeting 195m Pt-BP to metastatic bone lesions will trigger local release of 195m Pt near cancer cells upon vigorous bone remodeling and osteolytic resorption at reduced pH.
Herein, we provide solid evidence to prove our previously made claim on bone tumor-targeting of 195m Pt-BP complexes [15]. In more detail, we confirm for the first time that 195m Pt-BP complexes are indeed specifically targeted to bone metastases because of their high metabolic activity. First, we establish a bone metastasis model by intratibial injection of human prostate cancer cells and validate this model by monitoring bone metabolic activity (using 99m Tc-methylenediphosphonate [ 99m Tc-MDP], micro-single-photon emission computed tomography/computed tomography [micro-SPECT/CT] imaging) and lesion formation (using ex-vivo high-resolution micro-CT imaging) (Fig. 1B). Second, we confirm specific accumulation of 195m Pt-BP (but not of 195m Pt-cisplatin) in these tibial metastatic lesions using micro-SPECT/CT imaging (Fig. 1C). To compare the therapeutic efficacy of radioactive 195m Pt-BP vs. 195m Pt-cisplatin vs. non-radioactive Pt-BP species, we perform immunostainings to detect DNA damage and apoptosis in tibial metastatic lesions [16]. Overall, we provide the first preclinical evidence for the radiotherapeutic efficacy of Auger therapy for the treatment of bone metastases using 195m Pt-BP complexes, which stresses the novelty and potential impact of our approach.

Study design
The objective of the study was to confirm the bone tumor-targeting potential of 195m Pt-BP complexes and localize Auger electron-emitting 195m Pt radionuclides specifically within intratibial bone metastases. To this end, we first validated our intratibial bone metastasis model using 99m Tc-MDP micro-SPECT/CT imaging and ex vivo high-resolution micro-CT imaging. Blinding was applied in housing. In addition, biotechnicians were also blinded upon administration of different treatments. The mice were monitored for behavioral and bodily changes daily, starting at day 0 to prevent excessive discomfort due to excessive tumor growth, swelling, loss of body weight, and/or abdominal circumference. The following humane endpoints were considered as indications for the need of euthanasia: a) body weight gain of >20% within 10 days caused by tumor growth; b) no water or food intake; c) weight loss of >15% per 1-2 days or >20% of initial body weight; d) serious circulatory or respiratory problems; e) behavioral changes (hyperactivity, passive, automutilation) and locomotion problems; f) morbidities such as impaired mobility, necrosis, or lethargy; and g) excessive tumor growth leading to bone fractures. Image acquisition and data analysis of the in vivo and ex vivo data were performed in a non-blinded manner.

In vivo models
All in vivo studies were conducted in accordance with ISO standards and the principles set forth by the Revised Dutch Act on Animal Experimentation. The in vivo experiments were approved by the institutional Animal Welfare Committee of the Radboud University Medical Center (Radboudumc), Nijmegen, the Netherlands. For the in vivo experiment, male BALB/cAnNRj-Foxn1 nu /Foxn1 nu (athymic nude) mice (Charles River), with an average weight of~25 g and an age of approximately 6-8 weeks, were housed in filter-topped cages (5 mice per cage) under nonsterile standard conditions provided with standard animal food and water ad libitum. The mice were allowed to adapt to laboratory conditions for 1 week before experimental use.
For intratibial injection, the mice were anesthetized via isoflurane inhalation. PC3-eGFP or MDA-MB-231-eGFP or PC3-luc (2 Â 10 5 ) cells suspended in 20 μl cold PBS (phosphate-buffered saline) was injected into the right tibia, whereas only cold PBS was injected into the contralateral left tibia of BALB/cAnNRj-Foxn1 nu /Foxn1 nu mice using 30G needles (BD Micro-Fine). Analgesia was provided with buprenorphine (0.5 mg/ml) once before the intratibial injection. For validation of the bone metastasis model, PC3-eGFP and MDA-MB-231-eGFP cell lines were used. After intratibial injection of PC3-eGFP or MDA-MB-231-eGFP, mice were randomized into three groups, corresponding to weeks 1, 3, and 5 (6-8 mice per group). For the biodistribution study, PC3-luc cell lines were used. After intratibial injection of PC3-luc, mice were followed biweekly using bioluminescence imaging to confirm tumor growth and 18 F-NaF PET (positron emission tomography) imaging to confirm tumor-induced formation of lesions. Mice bearing tibial lesions were equally and randomly distributed by a biotechnician into four groups (n ¼ 4) to be treated with 195m Pt-BP, 195m Pt-cisplatin, Pt-BP, or saline control. 195m Pt/Pt complexes were administered intravenously (single dose) 1 day after randomization, followed by euthanasia after 2 weeks or at a humane endpoint. acquired using a U-SPECT-II/CT (MILabs), as reported previously [15,17,18]. The following criteria were used to confirm the tibial lesion formation: a) visual confirmation of 99m Tc-uptake below the growth plate in right tibia compared to contralateral control tibia; b) at least 10% 99m Tc-uptake increase in tibial lesion region of interest (ROI) below growth plate compared to contralateral control tibia.

SPECT imaging
Mice were scanned using a U-SPECT-II/CT (MILabs) under general anesthesia (isoflurane/O 2 ) for 15 min using the 1.0-mm diameter pinhole mouse high sensitivity collimator tube, followed by a CT scan (spatial resolution 160 mm, 65 kV, 615 mA) for anatomical reference. Scans were reconstructed with MILabs reconstruction software using an orderedsubset expectation-maximization algorithm, with a voxel size of 0.4 mm. SPECT/CT scans were analyzed and maximum intensity projections were created using the Inveon Research Workplace software (IRW, version 4.1). A three-dimensional (3D) volume of interest (VOI) was drawn using CT threshold (CT value: soft tissue is 11-39% and skeletal tissue is 41-100%) to differentiate soft tissue from skeletal tissue, and uptake was quantified as the percentage injected dose per gram (%ID/g) using standard curve, assuming a tissue density of 1 g/cm 3 . The hot spot  18 F-NaF PET imaging and bioluminescence imaging. Mice bearing tibial lesions were randomized into four treatment groups: 195m Pt-BP, 195m Pt-cisplatin, Pt-BP, and saline control. 195m Pt biodistribution was followed by micro-SPECT/CT imaging in mice bearing tibial lesions at 1 h, 24 h, and 7 days after administration. Tibial lesion and contralateral control tibia were analyzed using ex-vivo high-resolution micro-CT and histological methods. (Images were created using Biorender.) in the skeletal tissue ROI was chosen with the location of the edge of the ROI contour representing 75% of maximum intensity. All mice were euthanized with CO 2 after micro-SPECT/CT imaging.

Validation of bone metastases using high-resolution X-ray CT (HR-microCT)
After harvesting, the tibias were fixed in a freshly prepared 4% paraformaldehyde solution for 1 day and transferred to 70% ethanol for storage. Subsequently, they were scanned at 5 μm resolutions using HR-microCT (Phoenix NanoTom S, GE Measurement and Control Solutions), as reported [19]. HR-microCT images were generated on the X-ray CT facility of the Department of Development and Regeneration of the KU Leuven, financed by the Hercules foundation. The source was equipped with a tungsten target and operated at 70 kV and 120 μA. An aluminum filter of 0.5 mm was applied to reduce beam hardening. A fast mode setting (i.e. exposure time 500 ms, frame averaging 1 and skip 0) was used and the scanning time was 8 min per sample. Images were analyzed using CTAn (Bruker MicroCT, Kontich, Belgium), as reported [19]. The reconstructed images were rotated in 3D and saved with the tibia growth plate aligned at 90 with the z-axis of the image stack (DataViewer software, v1.5.6.2, SkyScan-Bruker). Trabecular and cortical bone parameters were not analyzed separately because of the massive destruction of the bone architecture. To determine bone loss, we selected 1,400 images (7 mm height) starting at 500 μm below the growth plate level. On this data set, a ROI was drawn manually, incorporating both the trabecular structure and the cortical bone. Using 3D analysis, bone volume fraction (BV/TV) and bone volume (BV) were calculated. Tibial lesion was considered to be established only if at least 5% change in total bone volume within ROI compared to contralateral tibia was observed. The selected ROI was visualized in 3D using CTVox (Bruker MicroCT, Kontich, Belgium).

Bioluminescence imaging
After intratibial injection of PC3-luc cells in mice right tibia, tumor growth in tibias was monitored by measuring luciferase activity using real-time in vivo imaging by means of an IVIS Lumina II system (Caliper Life Sciences, Hopkinton, MA), as described previously [20]. In brief, 200 μl D-luciferin (150 mg/kg of mice body weight; PerkinElmer, The Netherlands) dissolved in PBS were injected subcutaneously 5 min before imaging the mice. After anesthetization by isoflurane, mice were imaged using a field of view of 12.5 cm, medium binning factor and an exposure time of 60 s. Luminescence intensity was visualized using the Living Image 4.5 software (Caliper Life Sciences), shown as rainbow plots with automatic scale bars per measurement and represented as radiance units (photons/second/cm 2 /steradian).

PET imaging
All mice injected with PC3-luc cells in the tibia (right leg) received an intravenous injection biweekly (W2 and W4) with a dose of approximately 9 MBq 18 F-NaF. PET imaging was performed by scanning four mice simultaneously per acquisition. Mice were imaged under general anesthesia (2-3% isoflurane/O 2 ) for 15 min at 1 hafter administration. Scans were reconstructed using Inveon Acquisition Workplace software with iterative 3D ordered-subsets expectation maximization using a maximum a priori algorithm with shifted Poisson distribution, with the following parameters: matrix 256 Â 256 Â 161, pixel size 0.4 Â 0.4 Â 0.8 mm, with a corresponding beta of 0.05 mm, as reported previously [21]. and [ 195m Pt(NO 3 ) 2 (en)] (en ¼ ethylenediamine; radionuclide purity > 95%) were kindly provided by NRG (Petten, the Netherlands). 195m Pt-BP was synthesized using [ 195m Pt(NO 3 ) 2 (en)] as precursor in a similar manner as described for non-radioactive Pt-BP (see Supplementary Materials). The pH of 195m Pt-cisplatin was neutralized to pH 7 using 1 M NaOH and reconstituted in a sterile saline solution. Characterization of 195m Pt-BP or 195m Pt-cisplatin via elemental analysis and Electrospray Ionization-Mass Spectrometry was not possible because of an insufficient amount of residual radioactive platinum compounds. Radionuclide purity was approximately similar, as previously reported [15,22]. 195m Pt-BP, 195m Pt-cisplatin, and Pt-BP were reconstituted in 200 μl sterile saline solution (0.9% NaCl) and administered intravenously in the tail vein of BALB/cAnNRj-Foxn1 nu /Foxn1 nu mice (n ¼ 4 per each platinum complex). Mice treated with 195m Pt-BP received an intravenous injection with a dose of 9.0 AE 0.1 MBq (3.4 mM). Mice treated with 195m Pt-cisplatin group received an intravenous injection with a dose of 5.2 AE 0.2 MBq (2.5 mM), ensuring Pt dose was below the Pt toxicity dose of 6 mg/kg for mice [23]. Based on previous results showing toxicity for radioactive cisplatin (but not for 195m Pt-BP) at doses >6 mg/kg, we used a 195m Pt-cisplatin dose below that threshold [15,22]. Inherently, the radioactive dose of 195m Pt-cisplatin (5.2 AE 0.2 MBq) was relatively half compared to that of 195m Pt-BP (9.0 AE 0.1 MBq). The mice treated with Pt-BP received an intravenous injection of Pt-BP (0.6 mg per mice) corresponding to the platinum concentration in the 195m Pt-BP group. Immediately after injection of 195m Pt-BP or 195m Pt-cisplatin, images were acquired using a U-SPECT-II/CT system (MILabs) at 1 h, 24 h, and 7 days as reported previously for 99m Tc-medronate [15]. Whole-body SPECT/CT imaging was performed at 1 h and 24 h, whereas only the lower part of mice including the tibias was scanned at day 7 because of reduced radioactivity in the ROI at this time point (<1%ID). SPECT/CT scans were analyzed as reported earlier for 99m Tc-medronate [15]. At the end of the experiment, mice were euthanized, tibias were harvested, and fixed in a 4% formalin buffer for 48 h.

Immunohistochemical and histochemical staining
After fixation, tibia samples from each group were decalcified in an EDTA (ethylenediaminetetraacetic acid) solution, dehydrated in a series of alcohol, and embedded in paraffin, as reported previously [24]. The specimens were sectioned at a thickness of 5 μm using a standard microtome (RM 2165; Leica). All sections were cut parallel to the long axis of the tibia. Sections were mounted in triplicate on a glass slide. For each tumor tibia, different stainings were performed on adjacent sections for better correlation. Hematoxylin and eosin (H&E) staining was performed as reported previously [25].
Immunostaining of γ-H2AX using rabbit anti-human γ-H2AX antibody (2577s, Cell Signaling Technology; diluted 1:400) was performed as previously described [26]. Paraffin sections were rehydrated in a series of decreasing concentrations of ethanol, and antigen was retrieved in sodium citrate buffer (pH 6.0) at 70 C for 10 min. Subsequently, slides were blocked with 10% normal donkey serum (NDS), and then incubated with the primary antibody overnight at 4 C. Slides were then treated with a biotin-conjugated secondary antibody (Chemicon, Temecula, USA) for 1 hat room temperature, followed by counterstaining with methyl green. Negative controls using 2% NDS instead of the primary antibody were generated in parallel to ensure that the staining was specific. In addition, PC3-luc cells cultured in vitro were used as negative controls, whereas X-ray (2 Gy dose; XRAD 320 ix; Precision XRT; N. Brandford, CT, USA) irradiated PC3-luc cells cultured in vitro were used as positive controls. Finally, the sections were dehydrated and mounted. Apoptotic tumor cells were evaluated by FragEL DNA fragmentation detection kit (QIA33-1EA, EMD Millipore) with the colorimetric TdT enzyme (Calbiochem), following the manufacturer's protocol. Counterstaining with methyl green was performed for immunohistochemical staining of γ-H2AX and FragEL DNA fragmentation detection to aid in the morphological evaluation. All stained sections were scanned using an Olympus slide scanner (Olympus Nederland B.V.). Images were digitally viewed and cropped at a 10Â zoom level using Olympus OlyVIA software (v2.9). The effect of 195m Pt on the induction of DNA double-strand breaks was evaluated by immunohistochemical staining of γ-H2AX within the tumor ROI. At 10Â objective magnification, three different sections (n ¼ 3 per section) per tumor tibia at a distance of 100 μm from each other were prepared from the tumor-bearing tibias. Four random microscopic images at 10Â were obtained in the ROI within each section. To quantify colorimetric DAB (3,3 0 -diaminobenzidine) signal area of γ-H2AX-immunostaining, images were analyzed using ImageJ (version 1.52p) as reported earlier [27][28][29]. The images were selected in ImageJ, and the ROI was drawn within the selected images to determine the total tumor area. This procedure was followed by color thresholding to manually filter the DAB-stained area (dark brown) within the tumor ROI. The area corresponding to DAB-stained regions was measured. In addition, negative controls in the same slide were also analyzed as described earlier for each sample image to detect and eliminate false-positive values. Finally, an area-based analysis was used to extract a percentage of γ-H2AX positive tumor cell areas within the tumor ROI from each image. Similarly, the apoptotic tumor cell area was determined by FragEL DNA fragmentation detection to extract a percentage of apoptosis-positive tumor cell areas within the tumor ROI from each image.

Statistical analysis
Data for all parameters are expressed as means AE standard deviation. The statistical analyses were performed using GraphPad Prism (version 6.0) software. Two-way analysis of variance (ANOVA) with a Bonferroni (multiple comparisons) posthoc test was used to determine the differences among the two groups at different time points. For histology data, one-way ANOVA with a Turkey (multiple comparisons) posthoc test was used. For all statistical analysis, a value of p was considered as significantly different if *p 0.05; **p 0.01; ***p 0.001; ****p 0.0001.

Establishment and validation of metastatic lesions in mouse tibia
Bone metastases were established in the tibia of male athymic nude mice by intratibial injection of breast cancer cells (MDA-MB-231) or prostate cancer cells (PC3). Breast cancer is known to induce osteolytic lesions, whereas prostate cancer induces mixed lesions for these specific cell lines [30,31]. Hence, the induction of osteolytic or mixed lesions was validated at weeks 1 (W1), 3 (W3), and 5 (W5) after intratibial injection. We quantitatively assessed bone metabolism by 99m Tc-MDP uptake in the cancer cell-injected tibia (tibial lesion; right leg) and the PBS-injected tibia (control tibia; left leg) using micro-SPECT/CT imaging at 1 hafter administration. Furthermore, ex-vivo high-resolution micro-CT imaging of the tibial lesion and control tibia was performed to quantify the change in BV and BV/TV (BV in the total VOI). The corresponding quantification of the establishment of tibial lesions is summarized in Table S1.
Micro-SPECT/CT imaging of 99m Tc-MDP confirmed that tibial lesions were successfully established, as evidenced by increased bone metabolism at W5 for both prostate and breast cancer cell-induced lesions. Breast cancer cells often induced additional distant metastases, especially in the spine at W1 and W3 (Fig. S1A). However, injected prostate cancer cells induced localized tibial lesions of high metabolic activity (Fig. S1B). At W5, 99m Tc-MDP uptake in both types of tibial lesions was significantly higher (breast cancer: p < 0.0042, prostate cancer: p < 0.0001) compared with the contralateral control tibia (Fig. 2A). However, bone metabolic activity was higher in lesions induced by prostate cancer cells (10.1 AE 1.8% ID/g) vs. breast cancer cells (3.8 AE 0.7% ID/g). The selective uptake of 99m Tc-MDP in tibial lesions at W5 (relative to lesion-free control tibias) was also higher in lesions induced by prostate vs. breast cancer cells (Fig. 2B). Bone volume fraction and total bone volume of breast cancer cell-induced tibial lesions decreased from W1 to W5 because of osteolytic destruction of cortical bone (Fig. 2C and D), whereas prostate cancer cell-induced lesions showed a constant bone volume fraction. Furthermore, both bone volume fraction (Fig. 2C, at W3 and W5) and total bone volume (Fig. 2D, at W5) were higher for lesions induced by prostate vs. breast cancer cells. Ex-vivo high-resolution micro-CT imaging confirmed that breast cancer cells predominantly induced osteolytic lesions, whereas prostate cancer cells induced osteosclerotic lesions at W3 and mixed osteosclerotic/osteolytic lesions at W5 (Fig. 2E and F).
In brief, we conclude that intratibial injection of breast cancer cells result in tibial lesions with osteolytic activity as well as additional distant metastases in other parts of the skeleton. In contrast, intratibial injection of prostate cancer cells induce localized and mixed osteosclerotic/ osteolytic tibial lesions of increased metabolic activity. Consequently, we selected the prostate cancer cell-induced intratibial bone metastasis model to evaluate the targeting and therapeutic potential of 195m Pt-BP complexes.

195m Pt-BP specifically accumulates within tibial bone metastatic lesions
Using the intratibial prostate cancer cell-induced bone metastasis model, lesion formation and tumor growth in tibial lesions were measured biweekly using 18 F-NaF PET imaging (Fig. S2) and bioluminescence imaging (Fig. S3), respectively. The uptake of 18 F-NaF corresponds to changes in bone remodeling, whereas prostate cancer cells expressing luciferase facilitate longitudinal monitoring of tumor confinement [32,33]. Mice bearing tibial lesions received one of four different treatments in a randomized manner (4 weeks after intratibial injection, n ¼ 4): 195m Pt-BP, 195m Pt-cisplatin, Pt-BP, and saline control. As previous results showed toxicity for radioactive cisplatin at doses >6 mg/kg (but not for 195m Pt-BP), we used a 195m Pt-cisplatin dose below this threshold [15,22]. Inherently, the radioactive dose of 195m Pt-cisplatin (5.2 AE 0.2 MBq) was lower compared with 195m Pt-BP (9.0 AE 0.1 MBq). Ex-vivo high-resolution micro-CT imaging confirmed the formation of mixed osteosclerotic/osteolytic lesions in all mice except for three mice (1Â 195m Pt-BP, 1Â Pt-BP, and 1Â saline control) (Fig. S4). The bone volume within the metastatic lesion area increased with >15% compared with the contralateral control tibia for all treatment groups (Table S2). Nevertheless, the bone volume fraction after W6 varied considerably between mice because of the massive destruction of normal bone architecture (Table S2). 195m Pt-BP showed rapid bone tumor-targeted accumulation of 195m Pt in tibial lesions already 1 hafter systemic administration, whereas skeletal uptake of 195m Pt-cisplatin was not observed for the entire study period (Fig. 3A and B). The uptake of 195m Pt-BP in metastatic lesions was pronounced after 1 h (7.9 AE 0.3%ID/g, 22 μg of Pt/g), slightly decreased to 6.1 AE 0.5%ID/g (17 μg of Pt/g) at 24 h, and remained almost constant until day 7 (5.9 AE 1.1%ID/g, 16.4 μg of Pt/g).
To determine selective uptake of 195m Pt in the metastatic tibia, 195m Pt uptake was normalized to uptake in the contralateral lesion-free control tibia. 195m Pt-BP uptake was selective considering the 2.8 (AE0.6)-fold (at 1 h and 24 h) to 3.3 (AE2.2)-fold (at day 7) increased uptake in tibial lesions (Fig. 4B). In contrast, 195m Pt-cisplatin showed similar 195m Pt uptake in both metastatic and lesion-free control tibias at 1 h and 24 h. Targeting of 195m Pt-BP for metastatic lesions was further confirmed by its higher uptake compared to 195m Pt-cisplatin at 1 h, 24 h(2.7-fold), and 7 days (7.3-fold). Summarizing, these results clearly confirm that the BP ligand of 195m Pt-BP facilitates targeted accumulation of 195m Pt in prostate cancer cell-induced tibial lesions.
By applying CT thresholding, we were able to discriminate between accumulation of 195m Pt in soft tissue surrounding the lesion vs. skeletal tissue of the lesion (Fig. 4C). 195m Pt-BP showed almost similar accumulation (range: 1.8-2.1 %ID/g; 5.0-5.8 μg Pt/g) in soft tissue surrounding the tibial lesion at all time points. Similarly, 195m Pt-cisplatin also showed constant accumulation levels in soft tissue surrounding the tibial lesion at all time points, albeit at significantly lower (p < 0.001) values (0.3 AE 0.3%ID/g, 0.3 μg of Pt/g). Generally, accumulation of 195m Pt-BP in soft tissue surrounding the tibial lesion was three-fold lower than uptake in the entire tibial lesion after 24 hand remained constant until day 7. Similar to 195m Pt accumulation in the tibial lesion, accumulation of 195m Pt-BP in soft tissue surrounding the tibial lesions was also selective (Fig. 4D). These results demonstrate that targeted accumulation of 195m Pt-BP in tibial lesions also increases the 195m Pt level in soft tissue surrounding these lesions by a factor 5.5 (at 1 h) and 6.2 (at 24 hand 7 days) compared to 195m Pt-cisplatin.
Furthermore, we analyzed the total uptake of 195m Pt uptake in the soft vs. skeletal tissues of the entire mouse to differentiate the 195m Pt uptake in non-targeted tissue compared to targeted tissue (Fig. 4E). 195m Pt-BP mainly accumulated in skeletal tissue (3.2 AE 0.7%ID/g, 8.9 μg of Pt/g) at 1 hafter injection, which reduced to 2.6 AE 0.5%ID/g (7.2 μg of Pt/g) at 24 h. 195m Pt-BP uptake was significantly lower (p < 0.01 at 1 h, p < 0.001 at 24 h) in soft tissue than in skeletal tissue (1.5 AE 0.5%ID/g, 4.2 μg of Pt/g at 1 hand 0.7 AE 0.1%ID/g, 2 μg of Pt/g at 24 h). In contrast, 195m Pt-cisplatin showed more uptake in soft tissue (4.8 AE 1.0%ID/g, 5.4 μg of Pt/g) compared with skeletal tissue (2.3 AE 0.5%ID/g, 2.6 μg of Pt/ g) at 1 hafter injection, and showed equal uptake in soft vs. skeletal tissue (1.7 AE 0.2%ID/g, 1.9 μg of Pt/g) at 24 h. Finally, we compared the uptake of 195m Pt in the skeletal tissue of metastatic tibial lesions vs. uptake in the entire skeleton of the mice (Fig. 4F). 195m Pt-BP showed two-fold higher accumulation in skeletal tissue (6.6 AE 0.9%ID/g, 18.4 μg of Pt/g at 1 hand 4.3 AE 0.8%ID/g, 12 μg of Pt/g at 24 h) than 195m Pt-cisplatin (3.4 AE 1.2%ID/g, 3.8 μg of Pt/g at 1 hand 2.1 AE 0.4%ID/g, 2.4 μg of Pt/g at 24 h). Strikingly, 195m Pt-BP accumulation (9.8 AE 0.7%ID/g, 27.2 μg of Pt/g at 1 hand 7.8 AE 0.5%ID/ g, 21.7 μg of Pt/g at 24 h) was much higher in the skeletal tissue of tibial lesions than in the entire skeleton of the mice at 1 h and 24 h. In contrast, 195m Pt-cisplatin showed equal levels of 195m Pt accumulation (4.1 AE 1.0% ID/g, 4.6 μg of Pt/g at 1 hand 2.7 AE 0.5%ID/g, 3 μg of Pt/g at 24 h) in the lesions vs. the entire skeleton.
In summary, our data clearly confirm that 195m Pt-BP specifically accumulates in the metastatic lesions for our prostate cancer cell-induced intratibial bone metastasis model.

Targeted delivery of 195m Pt promotes radiation-induced DNA damage and apoptosis of metastatic tumor cells
To differentiate between bone marrow (dark purple) and tumor regions (light purple or purplish-pink), whole tibia sections treated with various types of Pt-based drugs were stained with H&E (14 days after start of treatment) (Fig. 5A-H). This staining confirmed the presence of tumor cell mass within the bone marrow and surrounding tibial lesions. One specific lesion treated with 195m Pt-BP showed the presence of a necrotic tumor region (Fig. 5D). Immunohistochemical staining of γ-H2AX, a biomarker specific for double-strand DNA breaks, confirmed DNA damage caused by either radiation or interstrand crosslinking of DNA by Pt-based drugs [16,34] ( Fig. 5I-L). Highest numbers of dark-stained γ-H2AX-positive tumor cells were observed upon 195m Pt-BP treatment (Fig. 5L). Moreover, apoptosis within the tumor region was examined by visualization of DNA fragments corresponding to apoptosis [35] (Fig. 5M-P). Similarly, the highest numbers of dark-stained apoptotic tumor cells within the tumor region were observed upon 195m Pt-BP treatment (Fig. 5P). Subsequently, we quantified the percentage of γ-H2AX positive and apoptotic areas within the tumor regions. The γ-H2AX-positive tumor cell area within the tumor region for mice treated with 195m Pt-BP (1.66 AE 0.4%) was 4.6-fold higher than 195m Pt-cisplatin (0.36 AE 0.1%), 11-fold higher than radio-inactive Pt-BP (0.15 AE 0.1%), and 32-fold higher than saline control (0.05 AE 0.04%) (Fig. 5Q). Similarly, the apoptotic tumor cell area within the tumor for mice treated with 195m Pt-BP (0.92 AE 0.5%) was 3.4-fold higher than treatment of mice with 195m Pt-cisplatin (0.27 AE 0.1%), 3.5-fold higher than radio-inactive Pt-BP (0.26 AE 0.2%), and 5.4-fold higher than saline control (0.17 AE 0.02%) (Fig. 5R). In conclusion, 195m Pt-BP treatment caused DNA damage and apoptosis in tumor cells more efficiently than all other treatment groups.
Finally, potential cytotoxic side-effects of 195m Pt on kidneys were evaluated histologically (Fig. S5). This analysis did not reveal morphological abnormalities following either 195m Pt-cisplatin or 195m Pt-BP treatment. Moreover, radiation-induced DNA damage or apoptosis were not observed in kidney tissue, as evidenced by the absence of positive γ-H2AX and apoptosis immunohistochemical staining.

Discussion
Radionuclides that combine abundant emission of low-energy Auger electrons with chemotherapeutic activity are highly attractive for cancer therapy because of their dual action against cancer cells. Therefore, previous in vitro and in vivo studies have explored the antitumor potential of Auger electron-emitting 195m Pt-radionuclides through the incorporation of 195m Pt in cisplatin [1]. However, undesired dose-limiting side-effects of 195m Pt-cisplatin (e.g. nephrotoxicity and ototoxicity) were comparable to radio-inactive cisplatin and caused by non-targeted tissue uptake [5]. Only few studies have evaluated the toxicity of 195m Pt associated with non-specific tissue or organ uptake [36,37]. Consequently, dose-limiting effects of clinically used cisplatin still hampers clinical translation of 195m Pt as a therapeutic agent. To target 195m Pt to bone metastases, the present study functionalized radioactive 195m Pt with a bone-targeting bisphosphonate ligand ( 195m Pt-BP). We demonstrated that BP ligands targeted 195m Pt effectively to metabolically active bone in metastatic lesions, while single systemic administration of 195m Pt-BP resulted in DNA damage and apoptosis of cancer cells.
This promising proof of concept could not have been obtained without a suitable bone metastasis model. However, to the best of our knowledge, standardized bone metastasis models are not yet available. Therefore, we established an intratibial bone metastasis model in mice to i) monitor 195m Pt accumulation in a localized bone metastasis relative to a contralateral control tibia in the same mouse; ii) ensure reproducibility with limited morbidity; and iii) minimize remote metastatic burden for the mice [38]. We induced this intratibial bone metastasis model using either human breast or prostate cancer cells and validated metastatic lesion formation based on bone metabolic activity ( 99m Tc-MDP; micro-SPECT/CT imaging) and alterations in bone morphology (mixed or osteolytic changes in tibial bone; ex-vivo high-resolution micro-CT imaging). Upon intratibial breast cancer cell injection, we observed osteolytic lesions, reduced bone metabolism, and formation of additional metastases, predominantly in the spine. In contrast, intratibial injection of prostate cancer cells resulted in the formation of a localized bone metastasis characterized by high bone metabolic activity and mixed osteosclerotic/osteolytic tibial lesions of high metabolic activity. Therefore, we selected the prostate cancer cell-induced intratibial metastasis model to evaluate the bone tumor-targeting and therapeutic efficacy of 195m Pt-BP vs. 195m Pt-cisplatin as control. Before evaluation of the bone tumor-targeting potential of 195m Pt-BP, successful establishment of intratibial bone metastasis was confirmed using 18 F-NaF PET imaging to avoid competitive inhibition between MDP and 195m Pt-BP. 18 F-NaF binds to bone by replacing hydroxyl groups, whereas bisphosphonates bind to calcium. Consequently, 18 F-NaF PET imaging was selected as imaging modality to confirm the formation of intratibial bone metastases [39]. Similar to our previous data on biodistribution of 195m Pt-BP [15], we herein observed specific accumulation of 195m Pt-BP in bone. Excitingly, accumulation in metastatic bone lesions was even higher already at 1 hafter administration compared with 195m Pt-cisplatin. This high accumulation in metastatic lesions can be attributed to the enhanced bone metabolic activity in bone lesions, as 195m Pt-BP accumulation was negligible in tibia with inferior lesion formation (<2% change in bone volume). In contrast, 195m Pt-cisplatin accumulated specifically in soft off-target tissues. However, none of the tumor-bearing mice showed weight loss, which contrasts recent findings by Aalbersberg et al. who used a 4.7 AE 0.2 MBqdose of 195m Pt-cisplatin in different strains of female mice [40]. Specific targeting of 195m Pt-BP in metastatic lesions resulted in the accumulation of~22 μg of Pt/g, while almost 75% of 195m Pt was retained in the metastatic lesions until day 7. Likewise, selective accumulation of 195m Pt-BP also increased the 195m Pt dose in soft tumor tissue surrounding tibial lesions. Thus, 195m Pt-BP uptake is accelerated by high bone metabolic activity in tibial lesions. Clinically, these findings are relevant since lesions of bone metastases in cancer patients are also characterized by high metabolic activity [41]. Moreover, 195m Pt-BP facilitated noninvasive detection of the tibial lesion within 1 hafter systemic administration, which opens up new avenues of research on theranostic application of 195m Pt-BP complexes.
To evaluate potential radiotherapeutic effects of 195m Pt-BP, we compared its therapeutic efficacy with non-radioactive Pt-BP and saline controls to study additional effects of Auger electrons emitted by 195m Pt. Previous work explored the therapeutic potential of Auger electrons emitted from, for example, 123 I, 125 I, and 111 In, which were conjugated to a monoclonal antibody or nucleus-specific peptide for targeting purposes [42][43][44]. As 195m Pt emits Auger electrons more efficiently, we functionalized 195m Pt with BP ligands to target these Auger emitters as close as possible to metastatic tibial lesions of high metabolic activity. Specific accumulation of 195m Pt enhanced DNA damage, as previously predicted by Monte Carlo simulations, which showed that 195m Pt radionuclides in close proximity to DNA induce cell death as caused by high-LET (linear energy transfer) α particles (5.3 MeV) [6,45]. Bone tumor-targeted delivery of 195m Pt-BP to metastatic tibial lesions damaged DNA in cancer cells 4.5-fold more efficiently than 195m Pt-cisplatin and even 11-fold more efficiently than non-radioactive Pt-BP. These results were affirmed by a significant increase in apoptotic tumor cells upon 195m Pt-BP treatment. Furthermore, one (out of 3) metastatic lesions in mice treated with 195m Pt-BP showed a necrotic region (~30%). Interestingly, this specific metastatic lesion also exhibited the most pronounced accumulation of 195m Pt-BP. Similar necrosis was also observed previously after 223 Ra (Xofigo) treatment in a patient-derived LuCaP 58 prostate cancer xenograft model [46]. These observations confirm that the therapeutic effect of low-LET Auger electrons at high concentrations is based on similar principles as conventional high-LET α-radiotherapy for treatment of bone metastases [1].
The present study highlights several advantages of 195m Pt-BP complexes for medical treatment of bone metastases based on their diagnostic and therapeutic potential. First, rapid uptake of 195m Pt-BP bone metastatic lesions confirms that 195m Pt-BP targeting is specific for bone tumors of high metabolic activity. These novel findings support the design of new theranostic agents which 'detect and treat' patients with bone metastases of high metabolic activity using Auger electron-emitting 195m Pt radionuclides. Second, 195m Pt-BP accumulation in bone metastatic lesions also increases 195m Pt distribution within soft tumor tissue surrounding metastatic lesions, which suggests that the increased DNA damage in the tumor region is caused by 195m Pt release into the metastatic lesion due to acidifying osteolytic effects. Third, targeting of 195m Pt-BP in metastatic lesions allows for higher dosing of 195m Pt compared with 195m Pt-cisplatin. Importantly, the targeted accumulation of 195m Pt-BP to metabolically active metastatic lesions also enhanced DNA damage to tumor cells 4.5-fold without causing nephrotoxicity. Finally, therapeutic anticancer effects of Pt-BP (expressed as DNA damage in tumor cells) increases 11-fold when using Auger-emitting 195m Pt instead of radio-inactive Pt. This radiotherapeutic efficacy of 195m Pt-BP was never shown before, which stresses the novelty and potential impact of our approach. Consequently, effective bone tumor-targeting of 195m Pt-BP paves the way for future clinical applications of 195m Pt-BP complexes as a theranostic agent to maximize therapeutic efficacy and minimize systemic toxicity.
Nevertheless, several limitations of the present study should be highlighted. For instance, a higher radioactive 195m Pt dose was used for 195m Pt-BP compared to 195m Pt-cisplatin for practical reasons explained in the experimental section. Proper follow-up studies are required to compare therapeutic effects at equal doses of both types of 195m Pt-drugs. Moreover, it should be emphasized that mice only received a single dose of Pt-based drugs at relatively low Pt concentrations. This experimental procedure might have compromised the chemotherapeutic efficacy of the Pt-based drugs as compared to multiple dosing regimens of clinically applied chemotherapy. Furthermore, the release of 195m Pt from 195m Pt-BP complexes was only confirmed indirectly by histological observations on tumor cell DNA damage and apoptosis. Although we previously showed in vivo release of 195m Pt from 195m Pt-BP by elemental mapping using laser ablation inductively coupled plasma mass spectrometry [15], immunohistochemical techniques to quantify tumor cell DNA damage and apoptosis were not compatible with this elemental mapping technique, which requires poly(methyl methacrylate) tissue embedding [47]. In addition, the localization of 195m Pt in bone marrow warrants further investigations to fully understand the effects of 195m Pt-BP on healthy bone cells. Finally, 195m Pt toxicity will still be a major concern for clinical translation, and hence needs further investigation using detailed dosing studies.
Therefore, further preclinical studies in larger animals and upscaling facilities are required before clinical translation becomes feasible. Specifically, these R&D efforts should focus on i) unraveling the relationship between therapeutic efficacy and specific radio-and chemotherapeutic activity of 195m Pt; ii) determining the minimum dosage required to effectively kill metastatic tumor cells while minimizing toxicity to surrounding healthy bone cells; and iii) upscaling the production of 195m Pt-BP complexes under good manufacturing practice conditions.

Data and materials availability
All data are provided in the paper or the supplementary section.