Focus Article

Development and applications of radioactive nanoparticles for imaging of biological systems Michael R. Lewis1,2∗ and Raghuraman Kannan3 Radioactive nanoparticles possess the ability to carry high payloads of radionuclides for noninvasive imaging of regions of interest inside the body. In this way, they can be used for nuclear imaging of systems such as normal physiology and disease states. Various methods have been developed to label nanoparticles using both radiometals and radiohalogens, for single-photon emission computed tomography (SPECT) or positron emission tomography (PET) imaging in laboratory animals. The use of imaging to develop radioactive nanoparticles with long circulation times and minimal reticuloendothelial uptakes led to the design of nanoparticle constructs for imaging animal models of chronic illnesses, such as cancer and cardiovascular disease. Further improvements in targeting were made by modifying these constructs with vectors having high affinity and specificity for diseased tissue. In addition, constructs containing more than one type of imaging material afforded nanoparticles with multimodal properties, such as those designed for nuclear, magnetic resonance, and/or optical detection. Given the close relationship between diagnosis and therapy, theranostic nanoparticles have also been developed both to deliver radiotherapy and monitor response by imaging. In this article, we review the use of radionuclides to label nanoparticles for development and applications involving noninvasive detection of normal and abnormal biological functions. © 2014 Wiley Periodicals, Inc. How to cite this article:

WIREs Nanomed Nanobiotechnol 2014, 6:628–640. doi: 10.1002/wnan.1292

INTRODUCTION

N

uclear imaging modalities include single-photon emission computed tomography (SPECT) and positron emission tomography (PET) as tools to visualize in vivo abnormalities. Nuclear imaging is noninvasive, and provides high sensitivity for detection of biological processes, especially in diagnosing and ∗ Correspondence

to: [email protected]

1 Research

Service, Harry S. Truman Memorial Veterans’ Hospital, Columbia, MO, USA 2 Department of Veterinary Medicine and Surgery, Nuclear Science and Engineering Institute, University of Missouri, Columbia, MO, USA 3 Departments of Radiology and Bioengineering, Center for Micro/Nano Systems and Nanotechnology, University of Missouri, Columbia, MO, USA Conflict of interest: The authors have declared no conflicts of interest for this article.

628

staging disease states. Several radiopharmaceuticals have been approved by the Food and Drug Administration (FDA) for use in humans. However, even at present, selective delivery of a radioisotope to visualize a particular region of interest (ROI) remains a challenge. In this context, nanoparticles (NPs) have emerged as promising vehicles to transport radioisotopes to desired sites within the body. Compared to small molecules, two potential advantages of NPs for in vivo applications are their ability to carry multiple copies of a targeting vector for increased binding avidity and their ability to carry large payloads of cargo for enhanced effector function. However, in order to take advantage of these properties, NPs need to have ‘stealth’ properties to avoid recognition by the mononuclear phagocyte system (MPS) and concomitant uptake by the reticuloendothelial system (RES), as well as relatively long blood circulation times. Clearance of particles through the

© 2014 Wiley Periodicals, Inc.

Volume 6, November/December 2014

WIREs Nanomedicine and Nanobiotechnology

Development and applications of radioactive nanoparticles

MPS occurs by a complex pathway of adsorption of serum proteins (i.e., opsonins) onto NPs, after which macrophage recognition, phagocytosis, and concomitant RES accumulation occur. Structure, surface modification, and charge, among other factors, can have considerable influence on this process. This focus article will discuss synthesis and applications of NP-based nuclear imaging agents with emphasis on their current status and challenges. Radioactive NPs for imaging can be designed by two possible methods (Figure 1). The first method (Type I) involves incorporation of a radioactive element into a nanosized cluster. Even though this method has excellent benefits, many radioactive elements tend to get oxidized at the nano level and subsequently elude the properties associated with nanoscale imaging. Noble metals such as gold can be bombarded with neutrons in a nuclear reactor to generate radioactive core NPs. The second method (Type II) involves attaching a radioactive element to a NP (also called radiolabeling of a particle). This method is versatile and can incorporate various radioelements of choice into a ligand on the NP surface. Typically, in this method multicarboxylate ligands are grafted on the surface of the NP. Subsequently, these chelators are used to carry metallic radionuclides. Even though this method is robust, the dissociation of the radionuclide under in vivo conditions could result in alteration of its in vivo distribution and false imaging. The properties of radioisotopes used for NP labeling by either method, typically used to evaluate in vivo stability and distribution, are given in Table 1.

RADIOACTIVE NP PREPARATION AND CHARACTERIZATION SPECT Isotope-labeled NPs Technetium-99m Technetium-99m is the most widely used SPECT radionuclide because it possesses optimal imaging characteristics, including a short half-life of 6.0 h and a 𝛾 emission of 140 keV for SPECT imaging applications. NPs have been labeled with 99m Tc to expand understanding of their biodistribution characteristics. Radiolabeling with 99m Tc has usually been accomplished using two different methods. In the case of iron oxide NPs, direct labeling with 99m Tc was performed by Madru et al.1 and Fu et al.2 However, other metallic or polymeric NPs have been modified by both the direct labeling approach and hydrazinonicotinic acid (HYNIC)-type ligand systems for labeling with 99m Tc. Volume 6, November/December 2014

In the case of direct labeling, superparamagnetic iron oxide nanoparticles (SPIONS) were treated with 99m TcO4 − in the presence of stannous chloride as the reducing agent.1 Using this method, 99% radiolabeling was achieved, as confirmed by conventional thin-layer chromatography. Technetium-99m-SPIONS showed homogenous size distribution with a mean diameter of 13 nm, as measured by transmission electron microscopy (TEM) studies. The zeta potential of the conjugate was between 5 and 15 mV at pH 4–6. In another study, CoFe2 O4 NPs were directly labeled with 99m Tc using SnCl2 .3 In this method, the conjugates were stabilized using ethyl 12-(hydroxyamino)-12-oxododeconate, poly(lactic-co-glycolic acid) (PLGA) and bovine serum albumin. The average hydrodynamic diameter of NPs increased from 106 to 160.8 nm after labeling with 99m Tc. The zeta potential remained constant (−23.6 to −23.8 mV) after labeling. High negative zeta potential of the nanoconjugate indicated that the product was stable towards aggregation. Zeta potential closer to zero indicates that a product is unstable and aggregates with time. Zeta potential measurements of NPs under normal conditions would be different from those observed in tissues. Under normal conditions, NP mobility is largely determined by hydrodynamic size and the potential created by the core (and attached surface molecules). Under conditions with high protein concentrations (such as in vitro or in vivo conditions), the NPs will be surrounded by a protein corona that would decrease the mobility and impact the zeta potential. Carrière et al.4 showed that the zeta potential of TiO2 shifts to a negative value when it enters in vivo systems, due to the formation of large clusters coated with numerous proteins. It is generally accepted that NPs attach to different proteins within an in vivo system, causing a major change in the zeta potential. Electron microscopy techniques, in conjunction with dynamic light scattering (DLS) measurements, were used to confirm the size of the NPs. It is important to realize that TEM provides the core size of the NP, whereas DLS provides the hydrodynamic size of the NP. In electron microscopy techniques, a carbon-based ligand framework that is used as a stabilizer is often not visible. Therefore, the TEM technique provides the size of the electron-rich central core. On the other hand, DLS measurements aid in understanding the overall size of the NP along with the stabilizer. DLS measurements provide the polydispersity index (PDI) of the NPs. Using the PDI, the homogeneity of the particles can be determined. The final product requires characterization by both

© 2014 Wiley Periodicals, Inc.

629

wires.wiley.com/nanomed

Focus Article

Stabilizing ligands

Radioactive isotope

Radioactive core

Core nanoparticles

NP*

Type I

Type II

FIGURE 1 | Schematic of Type I and Type II nanoparticle construction.

TEM and DLS techniques to unambiguously assign the size and dispersity of the final nanoconjugate. The study described above demonstrates that direct labeling of SPIONs with 99m TcO4 using a SnCl2 reducing agent yields high radiolabeling efficiency, and this technique is easy to translate into clinical settings. Increase in the hydrodynamic size of SPIONS after radiolabeling may serve as an optimal quality control to ascertain product formation. It is important to note that the direct radiolabeling technique is robust and can be utilized for both metal and polymeric nanoconjugates. In another report, the surface of 20-nm gold NPs was functionalized with both HYNIC-Gly-Gly-Cys-NH2 (HYNIC-GGC) and thiol-mannose.5 In this construct, the HYNIC ligand was conjugated with Gly-Gly-Cys. The thiol group present in the Cys terminus was attached to the gold NP. In addition, thiol-containing mannose was used as a targeting group to selectively accumulate the NPs in lymph node macrophages. TEM was used to monitor the core size of the NP before and after conjugation with HYNIC and mannose ligands. In fact, TEM showed low electron density close to the surface of the NP indicating the presence of organic ligands. Additionally, hydrodynamic diameter measurement indicated an increase in the size of the

NP before and after conjugation with organic ligands. Additional evidence for conjugation of both ligands to the surface of the NP was obtained from X-ray photoelectron spectroscopy (XPS) measurements. Indeed, XPS measurements indicated covalent interaction of thiol group in mannose and weak interaction of HYNIC ligand system (through –SH and –NH2 ). Radiolabeling with 99m Tc yielded the nanoconjugate, AuNP-mannose-HYNIC-99m Tc, in >95% purity. In comparison with the direct labeling approach, ligand-mediated conjugation of 99m Tc to NPs yields a well-defined product. The ligand-mediated approach is versatile and can be utilized with a wide variety of 99m Tc precursors. In contrast, the direct labeling approach is limited to the use of TcO4 − and a concomitant reduction of the radiometal.

Indium-111 As reviewed by Psimadas et al., 111 In-labeled NPs have been widely used to understand the biodistribution of NPs.6 For this purpose, polymeric and micellar NPs have been labeled with 111 In. Also, NPs of carbon, gold, or iron oxide were labeled with 111 In for understanding the biodistribution or evaluating disease states. Diethylenetriaminepentaacetic acid (DTPA) was used for 111 In labeling of gold and carbon

TABLE 1 Examples of Radionuclides, Nanoparticles, and Applications Radionuclide 99m 111

𝛾

𝛽+

𝛽−

Application

Example Nanoparticles

Objectives

Tc

X

SPECT

Colloid, iron oxide

Improve sensitivity and contrast

In

X

SPECT

Organic, inorganic

Develop SPECT or MRI agents

18

F

X

PET

Iron oxide and optical

Develop PET/MRI/optical agents

64

Cu

X

PET

Copolymers, iron oxide

Optimize biodistribution, PET, MRI

13

N

76

Br

X

X

PET

Directly activated

Develop in situ radiolabeling

X

PET

Targeted dendrimers

Selectively detect ischemia

153

Sm

X

X

Theranostic

Upconversion

Develop optical and therapy

198

Au

X

X

Theranostic

Gold

Develop SPECT and therapy

630

© 2014 Wiley Periodicals, Inc.

Volume 6, November/December 2014

WIREs Nanomedicine and Nanobiotechnology

Development and applications of radioactive nanoparticles

NPs, whereas iron oxide NPs were conjugated with 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) to build the ligand system.

PET Isotope-labeled NPs Copper-64 Superparamagnetic iron oxide (SPIO) NPs have been functionalized with DOTA and radiolabeled with 64 Cu.7 In this study, micelle-coated SPIOs were synthesized by treating 6.2-nm sized SPIOs with polyethylene glycol (PEG) derivatives, amino-PEG2000 and mPEG5000 phospholipids. Alternatives for PEG are evolving due to toxicity issues associated with extensive use of PEG in injectable formulations. The amino groups on the lipids were utilized to attach DOTA ligands to the surface. After coating with phospholipids, the hydrodynamic size of SPIOs increased from 7 to 20 nm. DOTA ligands on the surface of SPIOs were subsequently used to radiolabel with 64 Cu. In this study, nonspecifically bound 64 Cu was removed by treating the mixture with an excess of ethylenediaminetetraacetic acid (EDTA). The resultant conjugate showed excellent stability for 24 h in mouse serum. Wooley and coworkers8 developed shell cross-linked knedel-like (SCK) NPs. The authors found that the methodology used to functionalize with PEG had a profound impact on the biodistribution of NPs. For example, introduction of PEG groups to preformed NPs resulted in poor reproducibility. Incorporation of PEG molecules into block copolymer precursors, before conversion into NPs, produced constructs with longer blood residence time and lower liver uptake. This strategy is called ‘pre-grafting’, and this procedure helps in attaching an accurate number of PEG molecules on the NPs. The DOTA lysine derivative was conjugated to the block copolymer by conventional amidation. After attaching a sufficient number of DOTA-PEG molecules, the polymer was converted into NPs. Detailed characterization of the nanoconjugates showed that there was an average of two DOTA molecules per chain. As expected, increasing the PEG content per chain did not affect the hydrodynamic diameter (∼20 nm). In contrast, increased PEG content decreased the zeta potential of the resultant conjugate. The nanoconjugates synthesized in the study were spherical in shape and exhibited relatively narrow polydispersity. As mentioned previously, a narrow PDI indicates that the particles are homogeneous. Polymeric NPs of hydrodynamic sizes of 10–20 nm were synthesized with a poly(methyl methacrylate) (PMMA) core and a PEG shell and surface-functionalized with DOTA. The hydrodynamic size of the nanoconstruct changed from 9.7 Volume 6, November/December 2014

to 20 nm upon changing the PEG molecules on the surface from 1.1 to 5.0 kDa. A backbone comb polymer with low polydispersity was synthesized by terpolymerization of methyl methacrylate (MMA), methacryloxysuccinimide (MASI), and polyethylene glycol monomethacrylate (PEGMA). Owing to their comb architecture, these polymers dissolve only in dimethyl sulfoxide (DMSO) and not in water, unlike other polymers. These NPs were synthesized by treating DMSO solutions of comb polymer with water, followed by removal of DMSO by dialysis. The final well-defined NP was characterized by conventional analytical techniques. Each macromolecule contained 20–25 active functional groups available for attaching DOTA. Small Angle Neutron Scattering (SANS) analysis showed that the PEG (shell) thickness was 2.5 nm, with a core radius of 3.9 nm. Finally, 64 Cu labeling of the NP was achieved by mixing 64 Cu in acetate buffer with DOTA-functionalized polymeric nanoconjugates. Nonspecifically bound 64 Cu to amide and ester groups was removed by treating the radiolabeled NPs with DTPA with subsequent removal of DTPA-64 Cu complexes by Centricon separation.

Fluorine-18 Weissleder and coworkers9 synthesized 18 F-labeled paramagnetic iron oxide NPs using ‘click’ chemistry, a condensation of azide and alkyne groups catalyzed by Cu(I). In this study, aminated-cross-linked dextran iron oxide NPs of an average size of 30 nm, containing 40 primary amines per particle, were used. As a first step, five primary amine groups in each NP were used to functionalize with a fluorescent dye, Vivotag-680. The remaining primary amino groups present on the surface of the iron oxide NPs were derivatized to an azide functional group. Subsequently, azide groups on the NPs were chemoselectively reacted with propargyl 18 F-PEG in the presence of catalytic amounts of 3 copper salts to obtain 18 F-labeled iron oxide NPs in 99% radiochemical purity within 120 min. In the absence of copper salts, however, 18 F labeling could not be achieved.

USE OF IMAGING IN NP DESIGN In the studies described below, numerous radioimaging studies of Type II NPs in rats and mice were performed to assess the effects of NP construction on biodistribution, using noninvasive techniques. These studies evaluated RES uptakes of radioactive NPs as well as their residence times in the blood pool, correlating the results with conventional biodistribution experiments. In one of the early studies, Welch, Wooley, and coworkers10 examined the relationship

© 2014 Wiley Periodicals, Inc.

631

wires.wiley.com/nanomed

Focus Article

between size, core composition, and surface addition of PEG on the in vivo distributions of SCK NPs. These NPs were radiolabeled with 64 Cu for assessing biodistributions and microPET imaging of SCKs, focusing on blood retention and liver, spleen, lung, and kidney uptake in Sprague-Dawley rats. At 10 min post-injection, the conventional biodistribution of small NPs showed higher blood uptakes than larger particles. In addition, small particles with glassy polystyrene cores showed greater than 50% retention in the blood at 1 h post-injection and longer retention out to 4 h compared to those with fluid-like poly(methyl acrylate) cores, despite being of similar size (18 and 24 nm, respectively). In general, PEGylation (5240 Da) of the particles also decreased blood clearance, while having little to no effect on lung, liver, spleen, and kidney accumulation. These results showed that internal chemical composition, surface modification, and particle rigidity dictated longer residence times in blood. MicroPET imaging of small, PEGylated polystyrene NPs in normal Balb/c mice revealed high and persistent liver uptake out to 24 h, but the heart was clearly visible out to that time point (Figure 2), demonstrating the presence of NPs in the blood pool. In another report, Pressly et al. also evaluated the effects of PEG molecular weight on biodistribution and microPET imaging of block copolymer NPs,11 again using 64 Cu as the radiotracer. Particles modified with 1.1 kDa PEG exhibited relatively fast blood clearance in rats (27% remaining) at 4 h post-injection, compared to those with 2.0 or 5.0 kDa PEG (60–65% remaining). From 4 to 48 h, the blood retention of the 5.0 kDa construct decreased slowly, but that of the 2.0 kDa NP was lower at later time points. Interestingly, the intermediate sized NP showed characteristics of both high and low molecular weight particles with longer blood retention at early time points but more rapid clearance thereafter. These results indicate that the NPs could be fine-tuned with regard to circulation lifetime by manipulating their surface architecture. Inversely proportional to blood residence was liver uptake for the three NPs, and differences in accretion in the spleen and lung were unremarkable. More rapid kidney clearance was seen for the 1.1 kDa PEG-modified NP, concomitant with higher urinary excretion. Quantitative analysis of 64 Cu microPET images of mice was generally consistent with the biodistribution data obtained from rats. Again, the 2.0 kDa PEG particle showed intermediate characteristics between the high and low molecular weight species. Subsequent studies by Sun et al.8 employed 64 Cu to evaluate the impact of methoxy terminated-PEG 632

grafting of NP surfaces on in vivo behavior, the idea being that the complete coverage of NPs with mPEG2000 would impart ‘stealth’ properties capable of evading MPS recognition. In comparison with NPs containing 0, 200, and 500 mPEGs, the biodistribution in rats of a NP with 1100 mPEGs (20% cross-linked throughout the shell domain) showed prolonged blood values out to 4 h [2.3% injected dose per gram of tissue (% ID/g)] and a relatively rapid clearance thereafter (0.2% ID/g at 48 h). When particles containing 1100 mPEGs were cross-linked to a higher extent (50%), making them more robust and substantially reducing surface charge, significantly higher blood retention and lower accumulation in liver and spleen were observed. These results were in accord with the findings of several others that NPs have lower rates of opsonization. MicroPET imaging of mice demonstrated that particles with 1100 mPEGs produced high contrast heart images, as well as liver uptake and GI clearance. However, the NP with higher cross-linking showed the highest spleen uptake, contradictory to the biodistributions in rats. This observation raises the question of whether macrophage receptor populations differ between species and if a fair comparison can be made. As it is the most commonly used PET radionuclide, there has also been considerable interest in developing 18 F-labeled Type II NPs. Compared to 64 Cu, 18 F provides a shorter half-life, increased detection efficiency, and lower radiation dose to the subject (if that is a concern). In 2009, Weissleder and colleagues9 reported the preparation and study of 18 F-CLIO (cross-linked dextran iron oxide) NPs for PET/CT imaging of the blood pool and compared their results to MR imaging using the same particles. The PET imaging sensitivity in agar phantoms was approximately 200 times that of magnetic resonance imaging (MRI). PET/CT imaging was then performed in Balb/c mice from 1 to 16 h post-injection. High residence time in the cardiac chambers was observed initially, which over time decreased and liver and spleen uptake was detected. These NPs showed little kidney uptake, demonstrating that urinary excretion was not the predominant route of elimination. These investigators concluded the measured circulatory half-life of 5.8 h, as well as consequential accumulation in the RES, indicated macrophage internalization and phagocytosis. The following year, Welch and coworkers evaluated the PET/MR imaging capabilities of a 64 Cu-labeled PEGylated iron oxide NP.7 Micelle-coated superparamagnetic iron oxide particles (mSPIOs) exhibited high blood uptake at 1 h (37.3% ID/g), but by 4 h the NPs had cleared from circulation by a factor of nearly 5. At the end of 1 h,

© 2014 Wiley Periodicals, Inc.

Volume 6, November/December 2014

WIREs Nanomedicine and Nanobiotechnology

Development and applications of radioactive nanoparticles

1h

4h

24h

Heart

Liver

FIGURE 2 | Coronal microPET images of 64 Cu-SCK (top) and 64 Cu-PEG-SCK (bottom) in normal Balb/c mice. (Reprinted with permission from Ref 10. Copyright 2005 American Chemical Society)

accretion in the RES organs had occurred to a substantial degree (liver, 10.1% ID/g, and spleen, 20.0% ID/g). These biodistribution data were largely in agreement with tissue uptakes detected by microPET. The heart and carotid artery had high amounts of radioactivity at 1 h, which decreased after 4 h. RES uptake, mostly in the liver, was evident at 4 h. MRI capability was assessed by phantom studies, in which T2 -weighted contrast was strong using 25 μg Fe/mL and intermediate at 10 μg Fe/mL. More recent developments have focused on the study of radioactive Type I NPs. The physicochemical characteristics of Type I NPs, being intrinsically radioactive, have less potential for the radiolabel to detach during biodistribution and create false positive imaging results. Pérez-Campaña et al. built 13 N-labeled NPs by direct activation of naturally abundant 16 O in an aluminum oxide matrix.12 The 16 O(p,𝛼)13 N reaction was performed in situ by proton bombardment of the NPs in a biomedical cyclotron. This method was substantially more practical than the 18 O(p,n)18 F process, due to the high cost of the 18 O target. This group also demonstrated that the activation of the NPs did not alter their size or physicochemical properties. PET distribution of the particles at 10 and 60 min post-injection showed that small particles (10–40 nm) accumulated in the lungs (

Development and applications of radioactive nanoparticles for imaging of biological systems.

Radioactive nanoparticles possess the ability to carry high payloads of radionuclides for noninvasive imaging of regions of interest inside the body. ...
2MB Sizes 3 Downloads 6 Views