Chelator-Free Labeling of Layered Double Hydroxide Nanoparticles for in Vivo PET Imaging

Layered double hydroxide (LDH) nanomaterial has emerged as a novel delivery agent for biomedical applications due to its unique structure and properties. However, in vivo positron emission tomography (PET) imaging with LDH nanoparticles has not been achieved. The aim of this study is to explore chelator-free labeling of LDH nanoparticles with radioisotopes for in vivo PET imaging. Bivalent cation 64Cu2+ and trivalent cation 44Sc3+ were found to readily label LDH nanoparticles with excellent labeling efficiency and stability, whereas tetravalent cation 89Zr4+ could not label LDH since it does not fit into the LDH crystal structure. PET imaging shows that prominent tumor uptake was achieved in 4T1 breast cancer with 64Cu-LDH-BSA via passive targeting alone (7.7 ± 0.1%ID/g at 16 h post-injection; n = 3). These results support that LDH is a versatile platform that can be labeled with various bivalent and trivalent radiometals without comprising the native properties, highly desirable for PET image-guided drug delivery.

To perform PET imaging, radiometals (e.g. 64 Cu, 44 Sc and 89 Zr) constitute a primary category of radioisotopes. The coordination of certain chelators is usually required for the stable labeling of radionuclides 14,15 . However, due to the uniqueness of each radionuclide, particular coordination chemistry is necessary, and hence selecting the best chelator and optimal labeling condition becomes an inevitable challenge 15 . On the contrary, chelator-free labeling, which eliminates the need of coordination with chelators, has been considered as a promising labeling approach to nanoparticle-based PET imaging 16  M 3+ ) into the brucite-like cationic layers 17 . Therefore, we hypothesized that different radionuclides could be incorporated into LDH nanoparticles through a chelator-free manner, providing a novel nanoplatform for versatile radiolabeling.
Recently, we developed a reliable method to produce Mg 2 Al-LDH nanoparticles, which have a well controlled size range (50-300 nm) and can be homogeneously dispersed in the aqueous suspension 18,19 . Because LDH nanoparticles are highly positively charged (zeta potential: 30-50 mV) 19 , surface modification can be easily accomplished with negatively charged protein (e.g. bovine serum albumin [BSA]) to prevent possible aggregation in physiological buffer or in vivo environment, as recently reported 20 . Therefore, in this study, we propose chelator-free labeling of BSA-modified Mg 2 Al-LDH nanoparticles by simply mixing with a variety of PET isotopes (e.g. 64 Cu, 44 Sc and 89 Zr), followed by evaluation of the labeling yield and stability in vitro. In vivo PET imaging was subsequently carried out for the first time to investigate the in vivo biodistribution of radiolabeled LDH nanoparticles. Rapid and persistent tumor uptake via passive targeting was observed in subcutaneous 4T1 murine breast cancer models.

Results
Physicochemical features of LDH nanoparticles. Mg 2 Al-LDH nanoparticles were synthesized via quick precipitation and subsequent hydrothermal treatment as reported previously 1,18,19 . As shown in the transmission electron microscopy (TEM) image, LDH nanoparticles are 2-dimensional nanomaterials in layered-sheet shape with size in the range of 100-150 nm (Fig. 1b). Dynamic light scattering (DLS) measurement showed that LDH nanoparticles have an average size of 110 nm (polydispersity index [PDI]: 0.18) in water and 380 nm (PDI: 0.44) in culture media (Fig. 1d). The obviously larger size and PDI of LDH in culture media indicate that the unmodified nanoparticles are not very colloidally stable in physiological environment, which restricts the application of LDH in radiolabeling and in vivo PET imaging.
To improve the stability of LDH in vitro and in vivo, negatively charged BSA was used to coat the positively charged LDH surface via electrostatic interactions 20 . After the coating, as-prepared LDH-BSA nanoparticles remained stable in phosphate buffered saline (PBS) for 7 days, whereas LDH without BSA coating precipitated in PBS when incubated for the same duration (Fig. 1c). DLS measurements showed that the average size of LDH-BSA slightly increased to 130 (PDI: 0.21) in water and 150 (PDI: 0.24) in culture media due to the integration of BSA (Fig. 1d), which indicates that LDH-BSA remained stable in both water and PBS, consistent with the observations reported 20  LDH nanoparticles without BSA coating and only BSA were also labeled to validate that LDH but not BSA plays a more important role in the labeling process. Thin-layered chromatography (TLC) was used to measure the labeling yield with different isotopes using 50 mM EDTA as the mobile phase, which can remove the unbound and unstably adsorbed isotopes, ensuring that the resulting labeling yield truly represents the capacity of LDH or BSA for chelator-free labeling. 64 Cu 2+ and 44 Sc 3+ were found to successfully labeled LDH and LDH-BSA but not BSA, since LDH allows incorporation of multiple bivalent and trivalent cations 17 . On the contrary, 89 Zr 4+ could not label LDH and LDH-BSA, possibly because it does not fit into the LDH crystal structure ( Fig. 2; supplementary Fig. 1-3). In detail, after 60 min incubation, 59.0% 64 Cu 2+ was incorporated into LDH, while 16.6% was incorporated into LDH-BSA (Fig. 2a,b). The decrease in the labeling capacity of LDH-BSA is possibly attributed to the presence BSA, which prevents the interaction between 64 Cu 2+ and LDH. In addition, the stable labeling of 64 Cu 2+ on BSA was minimal (~1.5%), indicating that chelator-free labeling of 64 Cu 2+ mainly takes place on LDH rather than BSA. Similar results were also obtained with 44 Sc 3+ . After 60 min incubation, 41.4% 44 Sc 3+ was incorporated in LDH-BSA, much higher than that reacting with BSA only (1.7%; Fig. 2c,d). The evaluation of the labeling yield of LDH only with 44 Sc 3+ was not Zr 4+ cannot efficiently label onto LDH via chelator-free mechanism. Interestingly, 89 Zr 4+ could attach onto BSA, which is possibly due to the high affinity of 89 Zr 4+ to anionic oxygen donors in BSA 21,22 . Of note, chelator-free labeling was very quick, as the labeling yield peaked at 15 min and remained stable at the later time points for all three isotopes ( Fig. 2b,d,f). It is of concern that different specific activities of 64 Cu 2+ , 44 Sc 3+ and 89 Zr 4+ could be responsible for different radiolabeling outcomes. However, considering that chelator-free labeling takes place on the surface lattice of LDH nanoparticles which in turn provides excessively abundant labeling sites, the different specific activities are not expected to have significant impact on the radiolabeling.
Labeling stability is another common concern for chelator-free labeling. Besides the usage of EDTA solution during TLC analysis to remove unstably adsorbed ions, the stability of 64 Cu-LDH and 64 Cu-LDH-BSA was also tested in PBS and complete mouse serum before administration in living animals. After incubation in PBS for 24 h, 84.2 ± 11.4% and 89.3 ± 1.6% 64 Cu 2+ were still stable on LDH and LDH-BSA, respectively (Fig. 3). Similarly, after incubation with complete mouse serum for 24 h, 91.1 ± 3.9% and 91.6 ± 3.1% 64 Cu 2+ were stable on LDH and LDH-BSA, respectively (Fig. 3). These data demonstrate successful and stable chelator-free labeling of LDH and LDH-BSA nanoparticles.
One concern with in vivo PET imaging with 64 Cu-LDH-BSA is the possible detachment of BSA from LDH nanoparticles in blood stream. To investigate the influence of possibly detached of BSA on nanoparticle pharmacokinetics and tumor uptake, PET imaging was also performed with 64 Cu-BSA, although the labeling yield of 64 Cu-BSA is much lower than that of 64 Cu-LDH-BSA. After intravenous injection of 64 Cu-BSA into 4T1 tumor-bearing mice, a significantly lower tumor uptake was observed (2.9 ± 0.3, 3.4 ± 0.1, 4.1 ± 0.5 and 4.0 ± 0.3%ID/g, at 0.5 h, 3 h, 16 h and 24 h p.i., respectively; n = 3; Fig. 4 and Fig. 5b,c; supplementary Table 2). Moreover, the background signal in muscle was higher (1.9 ± 0.2, 1.5 ± 0.2, 1.1 ± 0.1 and 1.0 ± 0.1%ID/g, at 0.5 h, 3 h, 16 h and 24 h p.i., respectively; n = 3; Fig. 5b;  Table 2), possibly due to the smaller size and slower clearance of 64 Cu-BSA compared with 64 Cu-LDH-BSA, which led to a poorer imaging contrast (tumor/muscle ratio:1.5 ± 0.2, 2.2 ± 0.3, 3.6 ± 0.5 and 3.8 ± 0.5, at 0.5 h, 3 h, 16 h and 24 h p.i., respectively; n = 3; Fig. 5d; supplementary Table 3). Combining these data and the fact that BSA has a significantly lower labeling yield than that of LDH, we could draw a preliminary conclusion that the prominent tumor uptake of 64 Cu-LDH-BSA was primarily due to LDH, and any possible detachment of BSA in vivo will not significantly influence the imaging results.
After the last scan at 24 h p.i., 4T1 tumor-bearing mice injected with 64 Cu-LDH-BSA and 64 Cu-BSA were sacrificed for biodistribution studies. The quantitative data of the uptake in tumor and other tissues in biodistribution studies matched well with the results of ROI analysis, which confirmed the accuracy of PET imaging and ROI analysis (Fig. 6).

Discussion
Compared with conventional chelator-based labeling, chelator-free labeling possesses several unique advantages. First, chelator-free labeling of nanoparticles is a versatile approach, which may be suitable for various radionuclides. Therefore, a single design or strategy can potentially be employed in the different clinical settings, as exemplified by our current study. Secondly, it is known that incorporation of chelators could sometimes alter the in vivo biodistribution and pharmacokinetics of nanoparticles, which may not accurately reflect the pharmacological behavior of unlabeled nanoparticles [23][24][25] . Chelator-free labeling which avoids the use of a chelating agent, therefore, can maintain the native pharmacokinetic profile of the nanoparticles. Thirdly, chelator-free labeling overcomes the limitation of specific activities of radionuclides. Unlike traditional labeling procedures which depend on a limited number of chelators conjugated to the nanoparticles, chelator-free labeling takes place directly on the surface (or within   the lattice sites) of the nanoparticles via ion-exchange or other mechanisms. Therefore, abundant labeling sites are available resulting in high labeling yields, even though the specific activity is relatively low. Fourthly, chelator-free labeling does not require functional groups on the surface of nanoparticles for conjugation of chelators. Therefore, all the surface functional groups on the nanoparticles can be employed for further functionalization with fluorescent agents, therapeutic agents or targeting ligands to generate multifunctional nanoplatforms. And lastly, the mechanism for chelator-free labeling could be very simple and effective, allowing the labeling of isotopes such as 72 As and 69 Ge 15,16,26 , which to date is not easily achievable with a traditional chelator-based approach. Both 72 As and 69 Ge proved very challenging to label via conventional techniques, with few reported successes in the literature 15,26 .
However, chelator-free labeling also has certain drawbacks. Firstly, the surface modification of nanoparticles can significantly affect the labeling yield, since the surface of nanoparticles directly interacts with the isotopes during labeling. For example, BSA coating decreased the labeling yield of LDH-BSA with 64 Cu in this study. In addition, the labeling stability through chelator-free manner varies from case to case, differing for each isotope-nanoparticle combination. Since it is still a relatively underexplored area of research, very few examples can be referred to. In our previous study on chelator-free labeling of *As, obvious bladder uptake was observed in PET images with *As-SPION, which indicates that the labeling was not stable in vivo 15 . However, in this study, no bladder uptake was observed in vivo, suggesting a great potential of LDH nanoparticles for chelator-free labeling with excellent stability.
In this proof-of-principle study, ion-exchange is believed to play an important role in chelator-free labeling of LDH nanoparticles 27 . Through interactions between LDH nanoparticles and radiometals, the radioisotopes (guest species) tend to exchange with surface LDH atoms (host species) or compensate for the defect of LDH 27,28 . This allows the incorporation of various divalent and trivalent metal cations but not tetravalent metal cations, since only divalent and trivalent cations can fit into the LDH crystal lattice ([M 2+ 1−x M 3+ x (OH) 2 ] x+ (A n− ) x/n ·mH 2 O). However, further studies are needed to better understand the mechanism of chelator-free labeling in LDH nanoparticles. The valence of radioisotopes might not be the only decisive factor for this chelator-free labeling. Other factors, such as labeling buffer, temperature, pH value, concentration, and particle size, might also influence the labeling results 28,29 .
For nanoparticle-related imaging, toxicity of the nanoparticles is always one of the major concerns. Of note, LDH is considered one of the inorganic nanoparticles with low toxicity 30 . The toxicity of LDH is usually dose and time dependent, and the shape, size and surface charge also play a role in nanoparticle toxicity in vitro and in vivo 27 . In addition, aggregation or agglomeration of LDH nanoparticles upon suspension in physiological fluids is another reason for possible toxicity 3,31 . In this respect, BSA was coated onto LDH via electrostatic interactions, which is a straightforward and efficient process. The stability of LDH-BSA was significantly increased to prevent the possible aggregation in vivo. Besides BSA, other molecules, such as polyethylene glycol (PEG) or Tween-80, have also been used to modify LDH nanoparticles, which exhibited reduced toxicity in comparison with uncoated ones as previously reported 32,33 . Although LDH-nanoparticles have been proven to be highly biocompatible in vitro and many approaches have been investigated to modify the LDH surface for reduced potential toxicity, a long-term in vivo toxicity test is still of significance and warranted, considering that some elements (e.g. Al) can possibly lead to side effects if taken above a certain quantity or in the presence of certain physiological environments 3,34-36 .
The liver uptake is also a major concern for nanoparticle-based PET imaging. After intravenous injection into tumor-bearing mice, the nanoparticles are rapidly sequestered from the blood and accumulated in reticuloendothelial system (RES) organs (e.g. liver and spleen), and then gradually excreted by hepatic pathway (bile to feces) 37 . Depending on the material, size and surface charge of the nanoparticles, the excretion process may take hours to weeks. In this study, hepatic clearance of 64 Cu-LDH-BSA was evidenced by the decreasing liver uptake over time and radioactivity detected in the feces of mice. Due to relatively high signal in liver, renally non-clearable nanoparticles are not suitable for imaging liver tumors and liver metastases, although nanoparticles possess other advantages such as multimodality and theranostic potential. To reduce the liver uptake, great efforts have been made to improve surface modification of nanoparticle-based imaging agents 38 , which will be also critical for broadening the in vivo applications of LDH nanoparticles.
In conclusion, we report the chelator-free labeling and in vivo PET imaging with LDH nanoparticles. Upon surface modification with BSA, bivalent cation 64 Cu 2+ and trivalent 44 Sc 3+ cation successfully labeled LDH nanoparticles with excellent labeling efficiency and stability. Prompt and persistent tumor uptake was also observed with 64 Cu-LDH-BSA via passive targeting. Without compromising the native properties (e.g. drug loading), LDH can potentially serve as a versatile platform for PET imaging and drug delivery.

Materials.
Chelex 100 resin (50-100 mesh) was acquired from Sigma-Aldrich (St. Louis, MO). 64 Cu was produced in a GE PETrace cyclotron using the 64 Ni(p,n) 64 Cu reaction, which has a specific activity of ~185 MBq/nmol at the end of bombardment. 44 Sc was produced in a GE PETrace cyclotron using the 44 Ca(p,n) 44 Sc reaction, with a specific activity of ~8.2 MBq/nmol at the end of bombardment. 89 Zr was produced in a GE PETrace cyclotron using the 89 Y(p,n) 89 Zr reaction, with a specific activity of ~20.0 MBq/nmol at the end of bombardment. Complete mouse serum was acquired from Jackson Scientific RepoRts | 5:16930 | DOI: 10.1038/srep16930 Immuno Research Laboratories (West Grove, PA). Water and all buffers were of Millipore grade and pretreated with Chelex 100 resin to ensure that the aqueous solution was free of heavy metals. PD-10 desalting columns were acquired from GE Healthcare (Piscataway, NJ). All other reaction buffers and chemicals were from Thermo Fisher Scientific.

Cell lines and animal models. 4T1 murine breast cancer cells were obtained from American Type
Culture Collection (ATCC, Manassas, VA) and cultured according to the supplier's instructions. When they reached ~80% confluence, the cells were collected for tumor implantation 39 . Four-to five-week-old female BALB/c mice (Harlan, Indianapolis, IN) were each injected with 2 × 10 6 4T1 cells in the flank to generate the 4T1 breast cancer model. The mice were used for in vivo experiments when the tumor diameter reached 6-8 mm. All animal studies were conducted under a protocol approved by the University of Wisconsin Institutional Animal Care and Use Committee.
Synthesis of LDH. Mg 2 Al-based LDH nanoparticles were synthesized using a quick precipitation and subsequent hydrothermal treatment as reported previously 1,18,19 . In brief, 3.0 mmol of MgCl 2 ·6H 2 O and 1.0 mmol of AlCl 3 ·6H 2 O were dissolved in 10 mL deionized water, and then rapidly added to a basic solution (40 mL) containing 6.0 mmol of NaOH and 0.6 mmol of Na 2 CO 3 within 5 s. As-prepared solution was then stirred in N 2 stream at room temperature for 10 min. Subsequently, the precipitate was collected and washed twice with deionized water via centrifugation, and re-dispersed in 15 mL of deionized water and placed in a 20 mL autoclave with Teflon linen. Lastly, hydrothermal treatment was carried out at 100 °C in an oven for 16 h to generate the final suspension of Mg 2 Al-LDH nanoparticles with the mass concentration of 14.0 mg/mL.  [39][40][41] . The diluted solutions were then mixed with 50 μ L of LDH (4.7 mg/mL), LDH-BSA (4.7 mg/mL based on LDH concentration) or only BSA (33.3 mg/ml, the same concentration as that in LDH-BSA) for chelator-free labeling. The labeling reaction was carried out at 37 °C for 60 min under constant shaking. The labeling yield was measured by TLC using 50 mM ethylenediaminetetraacetic acid (EDTA) as the mobile phase to eliminate unstable adsorption of isotopes. The labeling yields at different reaction time were calculated from autoradiography images of TLC plates. The radioactive fraction was purified with PD-10 size exclusion column chromatography using PBS as the mobile phase.
To test the labeling stability, 64 Cu-LDH and 64 Cu-LDH-BSA were incubated in both PBS and complete mouse serum at 3 °C for up to 24 h under constant shaking. Portions of the mixture were sampled at different time points and filtered through 100 kDa MWCO filters. The radioactivity that remained on the filter was measured after discarding the filtrate, and retained (i.e., intact) 64 Cu-LDH or 64 Cu-LDH-BSA was calculated using the equation (radioactivity on filter/total sampling radioactivity × 100%). Cu-LDH-BSA via tail vein. Data acquisition, image reconstruction, and region-of-interest (ROI) analysis of the PET data were performed as previously described 13,42,43 . Quantitative data of ROI analysis on tumor and other organs were presented as percentage injected dose per gram of tissue (%ID/g). To validate the labeling and imaging of 64 Cu-LDH-BSA, 64 Cu-BSA was also scanned in 4T1 tumor-bearing mice (n = 3) as a control group. After the last scan at 24 h p.i., mice were sacrificed under anesthesia for biodistribution studies. Tumor, blood and major organs/tissues were collected and weighted. The radioactivity in the tissue was measured using a γ counter (PerkinElmer) and presented as %ID/g (mean ± SD).