Osteotropic Radiolabeled Nanophotosensitizer for Imaging and Treating Multiple Myeloma

Abstract

Rapid liver and spleen opsonization of systemically administered nanoparticles (NPs) for in vivo applications remains the Achilles’ heel of nanomedicine, allowing only a small fraction of the materials to reach the intended target tissue. Although focusing on diseases that reside in the natural disposal organs for nanoparticles is a viable option, it limits the plurality of lesions that could benefit from nanomedical interventions. Here we designed a theranostic nanoplatform consisting of reactive oxygen (ROS)-generating titanium dioxide (TiO₂) NPs, coated with a tumor-targeting agent, transferrin (Tf), and radiolabeled with a radionuclide (⁸⁹Zr) for targeting bone marrow, imaging the distribution of the NPs, and stimulating ROS generation for cell killing. Radiolabeling of TiO₂ NPs with ⁸⁹Zr afforded thermodynamically and kinetically stable chelate-free ⁸⁹Zr-TiO₂-Tf NPs without altering the NP morphology. Treatment of multiple myeloma (MM) cells, a disease of plasma cells originating in the bone marrow, with ⁸⁹Zr-TiO₂-Tf generated cytotoxic ROS to induce cancer cell killing via apoptosis pathway. Positron emission tomography/X-ray computed tomography (PET/CT) imaging and tissue biodistribution studies revealed that in vivo administration of ⁸⁹Zr-TiO₂-Tf in mice leveraged the osteotropic effect of ⁸⁹Zr to selectively localize about 70% of the injected radioactivity in mouse bone tissue. A combination of small animal PET/CT imaging of NP distribution and bioluminescence imaging of cancer progression showed that a single dose ⁸⁹Zr-TiO₂-Tf treatment in a disseminated MM mouse model completely inhibited cancer growth at euthanasia of untreated mice and at least doubled the survival of treated mice. Treatment of the mice with cold Zr-TiO₂-Tf, ⁸⁹Zr-oxalate, or ⁸⁹Zr-Tf had no therapeutic benefit compared to untreated controls. This study reveals an effective radionuclide sensitizing nanophototherapy paradigm for the treatment of MM and possibly other bone-associated malignancies.

The promise of nanomedicine in cancer therapy resides in the ability to deliver high payloads of drugs to cancer, amplify imaging signals,^1–5 optimize combination therapies, and improve theranostic strategies.^6–8 As vectors for drug delivery, nanoparticles (NPs) have the coveted advantages of modifying drug pharmacokinetics in vivo,^9–14 controlling drug release,^15–18 optimizing blood circulation half-lives,⁹ improving biodistribution profiles,¹⁹ increasing tissue permeability,^20–23 and enhancing metabolic stability of drugs.^{24, 25} For imaging applications, some NPs exhibit intrinsic properties that facilitate their use as imaging agents.^26–30 Typical examples include light-emitting quantum dots for fluorescence imaging and ultra-small superparamagnetic iron oxide particles as magnetic resonance imaging (MRI) contrast agents.³¹ Despite these benefits, clinical translation of most nanoparticle-based therapies and imaging agents has remained a challenge, in part due to the low percentage of injected dose that reaches the target tissues. This problem is particularly prevalent in metal-based nanoparticles that rapidly clear from circulation via liver, kidney, spleen or lung uptake. While some studies have leveraged this natural distribution pattern to deliver drugs and image the clearance organs, the lack of biological control limits the effectiveness and confines the disease type investigated. As a result, a large amount of disease-targeted NPs is needed to achieve a therapeutic effect, creating uneasiness about potential long-term toxicity.

Other than the aforementioned NP clearance organs, bone is an easily attainable tissue for targeted NP delivery. Coating nanoparticles with calcium-chelating ligands such as phosphonates is a viable option, but this does not alleviate the non-specific distribution in the liver and spleen. Previous studies have shown that some metals naturally home to the bone.³² This feature is used effectively to design radiopharmaceuticals such as ⁸⁹Sr, ¹⁵³Sm, ¹⁸⁶Re, and ²²³Ra for treating bone diseases and palliating pain.^{33, 34} By providing localized radiation, these agents can selectively ablate proliferating osteoclasts induced by cancer cells. However, the high dose of radioactivity employed can have significant side effects and in most cases, relapse is a frequent occurrence. Instead of directly killing cells with energetic particles, low dose positron-emitting radionuclides could be used to sensitize the therapeutic effects of drugs in the bone.

We recently reported the use of low-dose radionuclides to potentiate the therapeutic effects of ROS-generating TiO₂ NPs.^35–37 We demonstrated that the approach is capable of inhibiting tumor growth with minimal side effects via sequential administration of the NPs first, followed by radionuclides. Among other mechanisms, we found that TiO₂ NPs can harvest the UV light from Cerenkov radiation of fluorine-18 fluorodeoxyglucose (¹⁸FDG) to stimulate ROS generation. However, the uptake of ¹⁸FDG in bone is only twice higher in disseminated tumor-bearing mice than in healthy mice. Recent studies have shown that some ⁸⁹Zr chelates can distribute to diverse organs in mice, including bone tissue.^{38, 39} Although the selective osteotropism of ⁸⁹Zr oxalate was lost when chelated with desferrioxamine (DFO)³⁹ or adsorbed on liposomal membranes,³⁸ these suggest that integrating Zr into metal-based NPs could retain osteotropism if stabilized on an NP platform.

In this study, we developed a chelate-free ⁸⁹Zr-TiO₂ NPs, which were coated with a tumor-targeting Tf to afford thermodynamically and kinetically stable ⁸⁹Zr-TiO₂-Tf NPs. We then demonstrated the enhancement of ROS generation of the NPs compared to equivalent ⁸⁹Zr oxalate in vitro. The ⁸⁹Zr-TiO₂-Tf exhibited a high selectivity for bone, depositing about 70% of the injected radioactivity to the tissue. Administration of ⁸⁹Zr-TiO₂-Tf in disseminated MM mouse model, a bone-associated tumor, inhibited cancer growth and doubled the survival of the treated mice compared to the time of euthanasia of the untreated cohort. These results reveal an approach for bone-associated cancer therapy and uncover a potential strategy to treat bone lesions.

RESULTS AND DISCUSSION

Chelate-free Radiolabeling and Transferrin Coating of TiO₂ NPs

A major goal of this study is to deliver radiolabeled NPs to bone tissue with minimal uptake in the traditional NP clearance organs such as the lungs, liver, kidney, and spleen. Although Zr has been shown to possess osteotropic properties, the selectivity for bone depends on the nature of the metal. Chelation with DFO, for example, abrogates this feature.³⁹ As such, we sought a chelate-free approach to incorporate ⁸⁹Zr into TiO₂ NPs. Previous reports suggest that chelate-free, heat-induced ⁸⁹Zr radiolabeling of metal oxide NPs could be achieved through direct chemisorption.^40–44 We also explored the use of Tf to further stabilize the adsorbed Zr, enhance dispersion in aqueous medium, and target Tf receptor-expressing tumor cells. Apo-transferrin is known to chelate several metals such as Fe and Ti.^{45, 46} The overexpression of Tf receptor on rapidly dividing cells is essential for recruiting Fe as a co-factor during DNA synthesis.⁴⁷ We recently demonstrated that TiO₂-coated Tf is capable of targeting HT1080 fibrosarcoma and A549 non-small cell lung carcinoma.⁴⁸ Postulating that a combination of Zr and Tf attachment to TiO₂ would retain the osteotropic properties of Zr and target cancer cells, we used a chelate-free, heat-induced method,⁴⁰ to directly label anatase TiO₂ NPs with ⁸⁹Zr and coated the radiolabeled ⁸⁹Zr-TiO₂ NPs with Tf (Scheme 1).

Scheme 1:

Illustration of ⁸⁹Zr radiolabeling of TiO₂ NP and ⁸⁹Zr-TiO₂-Tf formation

The ⁸⁹Zr radiolabeling of the TiO₂ NPs was confirmed by radio-TLC using 50 mM DTPA aqueous eluent at pH 7.5 (Figure 1A). The final radiochemical yield of purified ⁸⁹Zr-TiO₂-Tf was >95% with specific activity of 18.5 MBq/mg. We further explored the efficiency of ⁸⁹Zr labeling by varying the TiO₂ NP concentrations, while maintaining ⁸⁹Zr activity at 9.25 MBq. We found that the radiochemical yield was > 98%, resulting in a specific activity of approximately 18.5 MBq/mg when radiolabeling 0.5 mg of TiO₂ NPs (Figure 1B). At lower concentrations of TiO₂ NPs, the radiochemical yield dropped significantly. Therefore, 0.5 mg of TiO₂ and 9.25 MBq ⁸⁹Zr were used for the NP construct preparation and subsequent studies. Overall, the heat-induced radiolabeling provided efficient chemisorption of ⁸⁹Zr on the TiO₂ NP surfaces, with high radiochemical purity and yield. The direct association of ⁸⁹Zr⁴⁺ with oxygen atoms on the surface of the multi-oxygen containing TiO₂ NPs^{49, 50} improved the radiochemical yield (RCY). The ⁸⁹Zr-TiO₂-Tf offers a simple integrated construct for imaging nanophotosensitizer in vivo and stimulating ROS for treating primary and secondary bone malignancies.

Figure 1:

Synthesis of ⁸⁹Zr-TiO₂-Tf (A): Radio-TLC profile of ⁸⁹Zr labeled TiO₂ challenged with 10 mM DTPA solution and (B): radiolabeling efficiency of ⁸⁹Zr-TiO₂-Tf with different amounts of TiO₂ NPs.

⁸⁹Zr-TiO₂-Tf is kinetically and thermodynamically stable

Although the radiolabeling experiments established that ⁸⁹Zr⁴⁺ metal ions attached to the surface of the TiO₂ NPs, this observation was not sufficient to confirm stable radiolabeling over time in vitro and in vivo. Therefore, we used ligand challenge and serum stability studies to assess the thermodynamic and kinetic stability of ⁸⁹Zr-TiO₂-Tf (Figure 2).^{51 89}Zr-TiO₂-Tf was stable in saline (Figure 2A) and 50 mM DTPA in water (Figure 2B), with > 99% of the ⁸⁹Zr radioactivity remaining bound to the high molecular weight NP fraction up to 72 h. In contrast, 72% of the ⁸⁹Zr radioactivity remained bound to TiO₂ NPs when challenged with DFO for 72 h (Figure 2C), due to its high affinity for Zr metal ions.⁵² Serum challenge data (Figure 2D) demonstrated that radiolabeled NPs were stable, with <10% of the ⁸⁹Zr radioactivity released as low MW material from ⁸⁹Zr-TiO₂-Tf. These data confirmed that chelate-free, heat-induced ⁸⁹Zr-radiolabeling generated a thermodynamically and kinetically stable product. In all cases, the radiolabeled NPs are stable within 24 h of incubation with diverse metal chelates. Although some radioactivity was lost at 72 h using DFO, the occurrence of this high-affinity Zr chelator in rodents and humans is so low that it is not a threat for the in vivo use of ⁸⁹Zr-TiO₂-Tf.

Figure 2:

Characterization of ⁸⁹Zr-TiO₂-Tf: Ligand challenge stability data acquired with size-exclusion chromatography at 0, 24, 48 and 72 h incubation at 37 °C in saline (A), 10 mM DTPA in water (B), 10 mM DFO in water (C), and mouse serum (D).

⁸⁹Zr labeling does not affect the morphology of TiO₂ NPs

After demonstrating the stable radiolabeling of ⁸⁹Zr-TiO₂-Tf, we performed non-radioactive studies to evaluate the chemical and morphological properties of the TiO₂ NPs before and after Zr chemisorption using transmission electron microscopy (TEM) and dynamic light scattering (DLS) analysis. DLS showed that the hydrodynamic diameter of the Zr-TiO₂-Tf was comparable to the TiO₂-Tf NPs, with an average diameter of 122 nm ± 16 nm and PDI 0.14 ± 0.04 over 5 preparation experiments of the nanoparticle (Figure 3A). TEM of the Zr-TiO₂-Tf showed that Zr labeling did not affect the morphology of the TiO₂-Tf (Figure 3C), which we previously reported.³⁵ Coupled Plasma-Optical Emission Spectrometer (ICP-OES) analysis showed there was no detectable free Zr in the supernatant after precipitation of the nanoparticles with centrifugation, confirmed the chemisorption of >99% of Zr ions on the TiO₂ NP surfaces.

Figure 3:

Characterization of ⁸⁹Zr-TiO₂-Tf: (A) DLS analysis of the cold Zr labeled TiO₂-Tf NPs size distribution. (B) Representative TEM images of cold Zr labeled TiO₂-Tf NPs; (C) Representative TEM image of TiO₂-Tf NPs.

⁸⁹Zr-TiO₂-Tf Potentiates ROS production in cells and cell-free media

Whereas therapeutic radiopharmaceuticals can generate ionizing ROS for cell killing via DNA damage, diagnostic radiotracers do not produce sufficient ROS for this purpose. In the presence of appropriate photosensitizers, some radionuclides at non-therapeutic doses can stimulate ROS production to exert their cytotoxic effects.^{35, 37, 53, 54} Unlike previous studies, we evaluated whether direct incorporation of a radionuclide into a nanophotosensitizer could potentiate ROS production. Using the ROS-sensor, H₂DCFDA dye to determine ROS production in a cell-free system, we found that ⁸⁹Zr-TiO₂-Tf generated significantly higher levels of ROS at all time points when compared to either ⁸⁹Zr alone or TiO₂-Tf NPs (Figure 4A). In particular, 72 h after particle formation, ⁸⁹Zr-TiO₂-Tf produced 30% and 60% more ROS than TiO₂-Tf and ⁸⁹Zr, respectively. The apparent ROS production by TiO₂-Tf NPs could be caused by γ-emissions from neighboring wells containing ⁸⁹Zr.

Figure 4:

In vitro properties of ⁸⁹Zr-TiO₂-Tf: (A) cell-free ROS production 24 h, 48 h, and 72 h after treatment with the indicated compounds; (B) MM1.S cell viability in response to different amounts of ⁸⁹Zr activity after treatment for 72 h; (C) ROS production by MM1.S cells 72 h after treatment with ⁸⁹Zr-TiO₂-Tf, ⁸⁹Zr, and TiO₂ compared to the untreated cells; (D) MM1.S cell viability 72 h after treatment with indicated compounds; (E) active caspase-3 levels in MM1.S cells 72 h after being treated with the indicated compounds. ⁸⁹Zr is used as ⁸⁹Zr-oxalate.

A critical tenet of radionuclide-augmented phototherapy, also known as Cerenkov radiation induced Therapy (CRIT) is the use of sub-toxic amounts of radiopharmaceuticals and photosensitizers to induce a therapeutic response. To establish the optimal dose of each component that will not induce cytotoxicity, we treated multiple myeloma (MM1.S) cells stably transfected with click beetle red (CBR) luciferase-mCherry construct and green fluorescent protein (GFP) with increasing amount of the ⁸⁹Zr and determined their viability 72 h later. The data showed that radioactivity higher than 0.37 MBq significantly affected the cell viability 72 h after treatment (Figure 4B), suggesting that this amount of activity is optimal for CRIT.

Although we demonstrated that ⁸⁹Zr-TiO₂-Tf generates ROS in a cell-free system, eukaryotic cells have developed effective anti-oxidative pathways and mechanisms to regulate oxidative stress,^{55, 56} which could prevent ⁸⁹Zr-TiO₂-Tf mediated ROS generation in cells. Treatment of MM cells with ⁸⁹Zr-TiO₂-Tf (0.37 MBq, 20 μg), followed by incubation for 72 h at 37 °C generated 25% more ROS than ⁸⁹Zr alone (0.37 MBq), 35% more than TiO₂-Tf (20 μg), and 38% more than the untreated cells (Figure 4C). These results demonstrate the potential use of ⁸⁹Zr-TiO₂-Tf for therapeutic applications.

⁸⁹Zr-TiO₂-Tf induces apoptosis of MM cells in vitro

We next evaluated the effect of ⁸⁹Zr-TiO₂-Tf on the viability of MM1.S cells in vitro. Treatment of MM1.S tumor cells with 0.37 MBq (20 μg) of ⁸⁹Zr-TiO₂-Tf decreased cell viability by about 50%, while neither ⁸⁹Zr alone nor TiO₂-Tf had any statistically significant effect (Figure 4D). The cell killing by ⁸⁹Zr-TiO₂-Tf could occur by multiple pathways. Apoptosis is one of the common downstream outcomes of prolonged exposure of cells to increased ROS levels.^57–59 Therefore, we evaluated the effect of ⁸⁹Zr-TiO₂-Tf exposure on the level of caspase-3 in MM1.S cells in vitro, a biomarker of caspase-mediated apoptosis. Indeed, treating MM1.S cells with 0.37 MBq (20 μg) ⁸⁹Zr-TiO₂-Tf for 72 h resulted in a 3.5-fold increase in activated caspase-3 compared to the untreated cells (Figure 4E). In contrast, ⁸⁹Zr, TiO₂-Tf, or a mixture of ⁸⁹Zr and TiO₂-Tf had minimal effect on caspase-3 expression. These results demonstrate that ⁸⁹Zr-TiO₂-Tf is capable of generating sufficient ROS to cause MM cell death, with caspase-dependent apoptosis playing a significant role.

⁸⁹Zr-TiO₂-Tf selectively localizes to bone marrow (BM)-containing bones

Positron emitting ⁸⁹Zr is an excellent radioisotope for PET imaging, allowing the imaging of ⁸⁹Zr-TiO₂-Tf distribution in vivo. To investigate the specificity of ⁸⁹Zr-TiO₂-Tf as an imaging agent, 1.11 MBq ⁸⁹Zr-TiO₂-Tf, ⁸⁹Zr-Tf, or ⁸⁹Zr-oxalate were injected intravenously (IV) into Balb/C mice (n = 5 per compound) and imaged with PET/X-rau computed tomography (CT) 48 h later. The imaging data revealed distinct localization patterns for the three compounds (Figure 5A). While ⁸⁹ Zr-oxalate localized primarily to the spine and ⁸⁹Zr-Tf instead localized primarily to the liver, ⁸⁹Zr-TiO₂-Tf accumulated predominantly in the spine, with moderate uptake observed in the liver. Quantification of the ⁸⁹Zr-TiO₂-Tf images gave standard uptake values (SUV) of 42 and 10 for all bone tissue and liver, respectively (Figure 5B). Ex vivo quantitative analysis of ⁸⁹Zr amount in excised tissue correlated with the noninvasive imaging data (Figure 5C). To confirm that ⁸⁹Zr-TiO₂-Tf was osteotropic and localized to the bones, we dissected the animals after the PET/CT imaging, harvested the organs, and quantified the amount of ⁸⁹Zr and Ti in each organ by gamma emission and ICP-OES, respectively. Both metals demonstrated excellent bone tropism, co-localizing to the bones and to a smaller extent in the liver (Figure 5D). In particular, the ⁸⁹Zr content in bones (42 %ID/g) was significantly higher than the amounts detected in any other organ (Figure 5C). Similarly, about 64 %ID/g of Ti was detected in the bones of animals treated with either ⁸⁹Zr-TiO₂-Tf or non-radiolabeled Zr-TiO₂-Tf (Figure 5D). Moreover, when the BM was separated from the matrix, it contained more than 50% of the total TiO₂ found in the bone (Figure 5E).

Figure 5:

In vivo properties of ⁸⁹Zr-TiO₂-Tf in Fox Chase SCID beige triple-immunodeficient mice (6 weeks old): (A) small animal PET/CT sagittal projections of images acquired 48 h after drug administration; (B) SUV for selected organs of interest obtained from images acquired 48 h after drug administration; (C) biodistribution of ⁸⁹Zr-TiO₂-Tf, ⁸⁹Zr-Tf, and ⁸⁹Zr-oxalate obtained with γ-emissions; (D) biodistribution comparison of Ti in Zr-TiO₂-Tf (blue) and ⁸⁹Zr-TiO₂-Tf (red) acquired with ICP-OES 168 h after administration of compounds (60 μg of each), p = 0.08 for bone uptake; (E) TiO₂ distribution between the BM and bone matrix acquired with ICP-OES.

Similar to the previous report,³⁹ chelation of ⁸⁹Zr with DFO in ⁸⁹Zr-Tf decreased the bone-seeking tendency of the metal. The dual distribution of ⁸⁹Zr-TiO₂-Tf in the bone and liver reflects the competing Zr osteotropism and Tf-driven liver uptake. Thus, the NP functionality can be used to modulate the distribution of Zr-labeled metal NPs. Although we focused on Zr osteotropism, previous studies with ⁴⁵Ti-oxalate (t_1/2 = 184.8 minutes), demonstrated the accumulation of this metal in the bone tissue as early as 5 h after administration.^{44, 60} This finding suggests that the chemical affinity of both Zr⁴⁺ and Ti⁴⁺ for the bone matrix could mediate or enhance the observed osteotropism. Taken together, the imaging and biodistribution results provided conclusive evidence that ⁸⁹Zr-TiO₂-Tf selectively localized to bone tissue, suggesting that it had the potential to provide image-guided functionality to MM CRIT.

Single-dose administration of ⁸⁹Zr-TiO₂-Tf inhibits MM progression in vivo

In addition to developing osteotropic nanophotosensitizers, we postulated that ⁸⁹Zr-induced ROS production could induce a therapeutic effect in a disseminated medullar MM xenograft. MM1.S cells, when administered IV in Fox Chase SCID beige mice, home to the BM and are readily detectable by a bioluminescence scanner 15–20 days later. After detecting the presence of the MM lesions in BM with bioluminescence imaging (BLI), we initiated ⁸⁹Zr-TiO₂-Tf therapy with a single IV injection of 1.11 MBq ⁸⁹Zr-TiO₂-Tf (60 μg TiO₂-Tf). We used PET/CT to monitor the distribution of the NPs (Figure 6A). As expected, the ⁸⁹Zr-TiO₂-Tf localized to the bones and liver and the BLI showed the engraftment of the tumors in bone tissues (Figure 6E). The tumor progression was monitored weekly with BLI. The single treatment with ⁸⁹Zr-TiO₂-Tf effectively suppressed tumor progression in the treated cohort compared to the untreated animals (Figure 6B). In separate experiments, we investigated whether the stable isotope of Zr-chloride, ⁸⁹Zr-oxalate, or ⁸⁹Zr-Tf would have any effect on tumor progression. These control animals received either 60 μg Zr-TiO₂-Tf (Figure 6C), 1.11 MBq (60 μg) ⁸⁹Zr-oxalate (Figure 6D), 1.11 MBq (60 μg) ⁸⁹Zr-Tf (Figure 6D), or no treatment and were monitored weekly with BLI. None of the control conditions had any appreciable effect on disease progression. These data clearly indicate that a single administration of ⁸⁹Zr-TiO₂-Tf resulted in specific and effective regression of MM disease progression in the treated cohort. In addition to suppressing the disease progression, animals treated with ⁸⁹Zr-TiO₂-Tf were still alive and presented 10-fold less tumor burden when all the untreated animals succumbed to the disease. This study demonstrates a successful application of single dose image-guided depth-independent radionuclide mediated nanophototherapy in a pre-clinical model of MM, facilitating the application of the technology in the treatment of diverse bone lesions.

Figure 6:

Therapeutic efficacy of ⁸⁹Zr-TiO₂-Tf: Therapeutic study of ⁸⁹Zr-TiO₂-Tf (1.11 MBq/60 μg) in SCID mice bearing MM1.S: (A) in vivo drug delivery tracking by PET 48 h after administration, (B) rate of tumor growth (whole body photon flux) in response to ⁸⁹Zr-TiO₂-Tf, (C) rate of tumor progression (whole body photon flux) in response to treatment with Zr-TiO₂-Tf, (D) rate of tumor progression in response to treatment with ⁸⁹Zr-oxalate and ⁸⁹Zr-Tf, (E) representative BLI images showing tumor localization 50 days after implantation.

Recent surge in the use of TiO₂ nanoparticles for medical applications has also raised concerns about their potential toxicity in living organisms.^61–65 Reports have shown that TiO₂ accumulates in the liver, spleen, lungs, and kidneys. The amount of organ uptake varied depending on the size and shape of the TiO₂ NPs, as well as the duration and route of particle administration. Some reports show that TiO₂ exposure induced the release of pro-inflammatory cytokines TNF-α, IL1-ß, and IL-6.^{66, 67} Unlike previous reports where TiO₂ doses of 40 mg/kg and above resulted in high liver uptake and various degrees of organ dysfunction, the small dose (3 mg/kg) used in our study exhibited moderate liver uptake of Ti and low accumulation in other organs (Figure 5).

CONCLUSION

We have established a method for synthesizing bone targeting ⁸⁹Zr-TiO₂-Tf NPs for single-dose image-guided CRIT of MM. Chelate-free, heat-induced radiolabeling of this construct was found to be thermodynamically and kinetically stable in vitro through a series of ligand challenge and plasma stability tests. The ⁸⁹Zr-TiO₂-Tf construct showed excellent ROS generation in cell-free and MM1.S cell model. Further, PET/CT imaging and biodistribution studies of treated mice showed that ⁸⁹Zr-TiO₂-Tf primarily localized to the bone, with the greatest accumulation in the BM. Previous studies have shown that the BM microenvironment serves as a niche for MM cells, providing a survival advantage and resistance to current MM treatments.⁶⁸ By targeting and infiltrating this niche, we showed that our treatment suppressed MM disease progression through continuous generation of the ROS from ⁸⁹Zr-TiO₂-Tf NPs, proportional to the half-life of the ⁸⁹Zr radionuclide. Additionally, these integrated multifunctional NPs provide insight into implementing CRIT for longitudinal therapy, while also monitoring therapeutic delivery using a single therapeutic dosage. Future studies will explore the use of this approach to deliver drugs to the BM, thereby minimizing systemic and multi-organ toxicity inherent in many chemotherapeutics.

MATERIALS AND METHODS

Anatase TiO₂ nanoparticles and all other chemicals, unless otherwise stated, were purchased from Sigma Aldrich (St. Louis, MO) and were used as received. Water (>18.2 MΩ·cm at 25 °C, Milli-Q, Millipore, Billerica, MA) was purified by passing it through a 10 cm long column of Chelex resin (Bio-Rad Laboratories, Hercules, CA) at a flow rate < 1.0 mL/min, which removed common metal ion contaminants. All instruments were calibrated and maintained in accordance with manufacturers’ instructions.

⁸⁹Zr Production and Purification

⁸⁹Zr was produced via the ⁸⁹Y(p,n)⁸⁹Zr reaction ay the Washington University Medical School Cyclotron Facility on a CS-15 cyclotron (Cyclotron Corporation, Berkeley, CA) and separated via ion exchange chromatography as previously described⁶⁹, with a resulting specific activity of 0.84–4.98MBq/μmole).

Radiochemistry

Radioactivity was measured with a Capintec CRC-15R Dose Calibrator (Capintec, Ramsey, NJ) with a calibration factor of 517 for ⁸⁹Zr. To accurately quantify radioactivity, experimental samples were analyzed for 1 minute on a calibrated Automatic Wizard2 Gamma Counter (Perkin Elmer, Waltham, MA), using a dynamic energy window of 800–1000 keV for ⁸⁹Zr (909 keV emission). Radiolabeling reactions were monitored with thin-layer chromatography (TLC) by using a radio-TLC plate reader (Lablogic, Brandon, FL). Solution pH was measured with pH paper strips (colorpHast non-bleeding strips; EMD Chemicals Inc., Gibbstown, NJ).

For labeling reactions, 10 mg/mL TiO₂ stock solution was prepared using commercial TiO₂ anatase powder suspended in metal-free water. TiO₂ NPs were radiolabeled with ⁸⁹Zr using heat-induced labeling without a chelator.⁴⁰ Briefly, an aqueous solution of ⁸⁹Zr (~ 37 MBq, approximately 0.1 mL, pH <1.0) in 1.0 M oxalic acid was transferred to a metal-free vial, and the solution was neutralized by adding 1 mL of 1.0 M 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer, pH 7.1. To optimize the radiolabeling conditions, aliquots of 9.25 MBq of ⁸⁹Zr were used for each reaction, with an amount of TiO₂ ranging from 10 μg to 160 μg. Reaction tubes were incubated at 100 °C for 1 h, with mixing at 1000 RPM. The reactions were then cooled to room temperature (RT). The radiochemical purity (RCP) was measured by radio-TLC. The radiolabeled NPs were combined with 6 mg apo-Tf in a w/w ratio of 1:3 Tf to TiO₂, followed by continuous probe sonication (Cole Parmer - Ultrasonic Processor GE 130PB, Vernon Hills, IL) at 3 watts for 1 minute. The final Tf coated radiolabeled NPs were filtered through a 0.22 μm polyethersulfone (PES) syringe filter (EMD Millipore, Burlington, MA) for the subsequent experiments.

To radiolabel Tf, Tf-DFO conjugate was prepared by mixing 10 mg/mL (0.012 μmol) Tf with 0.18 μmol 1-(4-isothiocyanatophenyl)-3-[6,17-dihydroxy-7,10,18,21-tetraoxo-27-(N-acetylhydroxylamino)- 6,11,17, 22- tetraazaheptaeicosine] thiourea (p-SCN-Bn-DFO, Macrocyclics, Plano, TX) using 0.1 M sodium carbonate (pH 9) as the conjugation buffer. The reaction mixture was vortexed at RT for 1 h. Then p-SCN-Bn-DFO was conjugated to Tf via a thiourea linkage and the conjugate was purified using a 0.5 mL Zeba spin column with a 40 kDa molecular weight cut-off (Thermo Fisher Scientific, St Peters, MO). The protein concentration of the resultant DFO functionalized Tf was determined by bicinchoninic acid assay (Thermo Fisher Scientific, St. Peters, MO). Tf-DFO conjugate (100 μg, 1.25 nmol) was then added to neutralized ⁸⁹Zr-oxalate (74 MBq in 1 mL sodium carbonate buffer, pH 9), and the reaction mixture was vortexed vigorously at RT for 1 h, followed by purification using Zeba spin columns. Radiochemical purity was determined by radio-TLC, using 50 mM diethylenetriaminepentaacetic acid (DTPA) as the mobile phase. The specific activity of ⁸⁹Zr-Tf-DFO was 55.5 MBq/nmol.

Stability

The stability of radiolabeled TiO₂ NPs was evaluated by performing ligand challenge experiments. Radiolabeled NPs (e.g., ⁸⁹Zr-TiO₂-Tf, 7.4 MBq, ~50 μl) were added to 450 μL solutions of either sterile saline, DFO (10 mM, pH7.0), or DTPA (50 mM, pH 7.5) and reactions were incubated at 37 °C for up to 72 h. Aliquots (15 μL) were removed at 0, 24, 48, and 72 h, and analysis of radiochemical stability was performed by using analytical size-exclusion chromatography (SEC) Sephadex G-25 (Thermo Fisher Scientific, St Peters, MO) columns, eluted with 16 × 0.5 mL fractions of solutions. Activity eluted in the first 2 mL was associated with the high molecular weight (>30 kDa) fraction, whereas activity eluted in the fractions >2.0 mL was associated with small-molecules derived from radiochemical decomposition under the ligand challenge conditions. The radiochemical stability of ⁸⁹Zr-TiO₂-Tf was also evaluated in normal mouse serum. Radiolabeled NPs (e.g. ⁸⁹Zr-TiO₂-Tf; 7.4 MBq; 50 μL) were incubated in either mouse or human serum (450 μL) for 72 h at 37 °C. Samples were analyzed by analytical SEC on PD-10 columns, eluted with 16 × 0.5 mL fractions of sterile saline.

Non-Radioactive Labeling Reactions

To facilitate a more detailed chemical analysis and morphology study of ⁸⁹Zr-labeled TiO₂ NPs, reactions were performed using non-radioactive Zr metal ion oxalate salts (⁹⁰Zr) evaluated for their potential to associate with TiO₂ NPs under chelate-free, heat-induced labeling conditions. Non-radioactive labeling reactions were conducted under the same conditions used in the radioactive studies with TiO₂ (0.5 mg) and Zr oxalate (0.55 mg, 2 μmol), followed by coating the labeled NP with 1.5 mg Tf. TEM and DLS were used to characterize the morphology and size of the ⁹⁰Zr-TiO₂-Tf NP construct. TEM was obtained using an FEI Tecnai Spirit microscope (Hillsboro, OR). The particle size distribution was measured by DLS using a Malvern Zetasizer Nano-ZS (Malvern Instruments Inc., Westborough, MA).

Reactive Oxygen Species Assay

TiO₂ NPs are known to generate ROS upon the irradiation by UV light or Cerenkov radiation. Using 0.37 MBq ⁸⁹Zr-TiO₂-Tf, ROS measurements were carried out in a Costar^@ 96- well black plate (Corning, NY) at RT with 2’, 7’-dichlorodihydrofluorescein diacetate (H₂DCFDA) as a ROS indicator. Measurements were performed using 0.37 MBq ⁸⁹Zr (same for both, ⁸⁹Zr-oxalate and ⁸⁹Zr-TiO2-Tf) and 20 μg TiO₂-Tf NPs. Emitted fluorescence was detected with a BioTeck Neo2 Hybrid Multi-Mode Reader (Winooski, VT) at 24, 48, and 72 h.

Cell Lines

MM1.S human multiple myeloma⁷⁰ cell line stably transfected with click beetle red (CBR) luciferase and green fluorescent protein (GFP) reporters was generously provided by Dr. DiPersio (Washington University School of Medicine).

In Vitro CRIT

In vitro CRIT experiments were performed in 24 well plates (TPP, Midwest Scientific, St. Louis, MO) containing 4×10⁵ MM1.S human multiple myeloma cells per well in 1 mL of complete medium (CM) (Iscove’s modified Dulbecco’s medium), supplemented with 10% fetal bovine serum and 50 μg/mL Gentamycin. Untreated cells or cells treated with ⁸⁹Zr-TiO₂-Tf NPs (0.37 MBq, 20 μg, in HEPES buffer), 0.37 MBq ⁸⁹Zr-oxalate, or 20 μg TiO₂-Tf were added to the wells and incubated at 37°C / 5% CO₂ for 72 h. The medium was then replaced with 500 μl PBS or 100 μl RIPA lysis buffer (Cell Signaling, Danvers, MA). Cell viability was measured with CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega, Madison, WI). Emitted fluorescence was detected with a BioTeck Neo2 Hybrid Multi-Mode Reader (Winooski, VT).

Caspase 3/7 Activity Assay

Caspase 3/7 activity were determined with Apo-ONE® Homogeneous Caspase-3/7 Assay kit (Promega, Madison, WI) according to the manufacturer’s instructions. Briefly, MM1.S cells were plated in 24 well plates at 4×10⁵ per well in CM followed by the treatment with ⁸⁹Zr-TiO₂-Tf NPs (0.37 MBq, 20 μg), 0.37 MBq ⁸⁹Zr-oxalate, or 20 μg TiO₂-Tf. The medium was replaced 72 h later with 250 μL PBS and 250 μL Apo-ONE® Homogeneous Caspase-3/7 reagent mixture was added to the wells and Emitted fluorescence was detected with a BioTeck Neo2 Hybrid Multi-Mode Reader (Winooski, VT).

Animal Care

All animal experiments were conducted in compliance with the Institutional Animal Care and Use Committee (IACUC) guidelines and the Guide for the Care and Use of Laboratory Animals. Fox Chase SCID beige triple-immunodeficient mice (6 weeks old) were obtained from Charles River Laboratory (Wilmington, MA). Animals were allowed to acclimatize for 1 week prior to studies, and were provided with food and water ad libitum. All compounds were suspended in 100 μL saline.

In Vivo CRIT

In vivo CRIT was performed in a disseminated xenograft model. Fox Chase SCID beige mice (n=5) were implanted with 2×10⁶ of MM1.S cells in saline through the tail vein. Eight days after implantation tumor progression monitoring was initiated and repeated weekly using BLI. The flux of 10⁶ photons/second was considered baseline, at which time a single dose of 1.11 MBq (60 μg) in 100 μL saline was administered.⁷¹ The tumor burden was monitored with weekly BLI.

Bioluminescence Imaging

In vivo BLI was performed on an IVIS Lumina (PerkinElmer, Waltham, MA; Living Image 3.2, 1–300 seconds exposures, binning 2–8, FOV 12.5cm, f/stop 1, and open filter). Mice were injected intraperitoneally with 150 mg/kg D-luciferin in PBS (Gold Biotechnology, St. Louis, MO) and imaged 10 minutes later under isoflurane anesthesia (2% vaporized in O₂). The total photon flux (photons/sec) was measured from regions of interest (ROIs) over the entire ventral side of the mice, using Living Image 2.6.

Small-Animal PET/CT Imaging

Prior to imaging, mice were injected intravenously via the tail vein with 1.11 MBq (60 μg TiO₂-Tf) of ⁸⁹Zr-TiO₂-Tf suspended in 100 μl saline. At 48 h post-injection, mice (n=3) were anesthetized with 1%−2% isoflurane and imaged with an Inveon small animal PET/CT scanner (Siemens Medical Solutions, Tarrytown, NY). Static images were collected for 20 minutes and reconstructed with the maximum aposteriori probability algorithm (MAP),⁷² followed by co-registration with Inveon Research Workstation (IRW) image display software (Siemens Medical Solutions, Knoxville, TN). ROIs were selected from PET images using CT anatomical guidelines, and the activity associated with them was measured with IRW software. SUV were determined as Bq/cc X animal weight/ injected dose.

In Vivo Biodistribution Studies

Biodistribution studies were conducted 168 h after ⁸⁹Zr-TiO₂-Tf treatment. Briefly, 1.11 MBq ⁸⁹Zr-TiO₂-Tf (60 μg) suspended in 100 μl saline was injected via the tail vein. Animals (n=4) were euthanized at selected time points after NP administration, organs of interest were harvested, weighed, and the associated radioactivity was determined with a γ-counter. After correcting for background and decay, the percent-injected dose per gram (%ID/gram) and percent-injected dose per organ (%ID/organ) were calculated by comparison to a weighed, counted standard. For non-radioactive biodistribution studies, ⁹⁰Zr-TiO₂-Tf (60 μg) suspended in 100 μL saline was injected in tail vein (n=4). To determine the relative amount of TiO₂ contained in the BM, the BM was removed by centrifugation at 10,000xg for 10 minutes and the TiO₂ content in BM and the remaining bone matrix was determined with induction-coupled plasma-mass spectrometry (ICP-OES).

Induction-coupled plasma-optical emission spectrometry

Mouse tissues were added to Teflon digestion vessels along with 2mL of concentrated nitric acid (70%, trace metal grade). The sealed digestion vessels were then placed into a Mars 6 Microwave Digestion System (CEM Corporation). Microwave power was ramped to reach 200°C in 20 minutes, then that temperature was maintained for 20 minutes. The sample was then cooled and diluted to 3 mL with 18MΩ DI water. A Perkin Elmer Optima 7300DV Inductively Coupled Plasma-Optical Emission Spectrometer (ICP-OES) was used to determine the Ti concentration of the diluted samples. Ti calibration standards of 1 ug/L, 10 ug/L, 100 ug/L, and 1000 ug/L were used. The internal standard (200ug/L, Sc) was continuously introduced during ICP-OES analysis. Ti concentration in the tissue was calculated by multiplying the measured Ti concentration by the dilution factor and the known sample volume then dividing by the tissue weight. Ti concentration was converted to TiO₂ concentration by multiplying by the molecular weight (M/W _TiO2 = 79.9 g/mol) to atomic weight (M/W_Ti = (47.9 g/mol) ratio.

Data Analysis and Statistics

Data and statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, Inc., La Jolla, CA) and Microsoft Excel. The data were expressed as mean ± SD, unless indicated otherwise. Differences at the 95% confidence level (P<0.05) were considered to be statistically significant.