Approaches to Nanoparticle Labeling: A Review of Fluorescent, Radiological, and Metallic Techniques

Abstract

Nanomaterials are widely used in commercial products, resulting in the release of nanoscale particles into the environment. This raises concerns about their potential exposure risks in complex biological matrices. Most attempts use engineered nanomaterials (ENMs) to mimic the biological behavior of nanoparticles in the environment, and labeling of ENMs by sensors is a commonly used approach for sensitive detection and tracking of ENMs in organisms. However, due to the distinct physicochemical properties of nanoparticles, different labeling approaches have been developed, each with varying applicability. In this Review, we summarize the three main types of labeling methods used for nanoparticles: fluorescent, radiological, and metallic labeling. We discuss their labeling mechanisms, efficiency, stability, target nanoparticles, and applicability in different organism models. Finally, we propose a labeling scheme for specific nanoparticles. Overall, this Review provides a comprehensive overview of the advances in nanoparticle labeling techniques and their potential applications in environmental and health studies.

1. Introduction

Nanotechnology has become an integral part of various industries, such as energy, electronics, mechanics, and healthcare. The global market currently boasts more than 10 000 commercial nanoproducts.¹ However, the extensive use of nanomaterials in commercial products has led to a significant release of nanoscale particles (e.g., carbon black, TiO₂, and SiO₂) into the environment, raising concerns about their potential health and environmental impacts. Detecting and tracking these invisible particles is crucial for their safety assessment. The released nanoparticles have been detected in water, soil, and air, with studies reporting their concentration in various environmental matrices. For instance, Gondikas et al. detected a mass concentration of 1.38 μg/L TiO₂ in the Old Danube Recreational Lake.² In another study, Zhang and Liu found plastic particles in all soil samples collected from different areas, displaying a concentration range from 7100 to 42 960 particles/kg.³ Shi et al. detected 53 500/cm³ atmospheric particles in the size range of 5.6–560 nm in an urban area of Beijing.⁴ These nanoparticles may be transported in different environmental media, accumulate in bacteria, plants, and animals, and cause adverse outcomes. For instance, metallic nanoparticles are often taken into cell lysosomes, eliciting pro-inflammatory cytokine release and cell death.⁵ Once deposited into the alveoli of animal lungs, the nanoparticles may induce lung inflammation,^6,7 fibrosis,^8,9 or carcinogenesis.^10,11 Understanding the biodistribution and fate of nanoparticles in the environment and biological systems is critical for their risk assessment and management.

Two prominent techniques for identifying targets in complex matrices are highly efficient separation and specific labeling. While some advanced separation methods, such as gas/liquid chromatography and electrophoresis, have been successful in separating small or biological molecules,^12,13 they are limited in separating nanoparticles. Therefore, chemical labeling is the most convincing and sensitive method for identifying nanoparticles in the environment and biological systems. Various labeling methods have been developed to detect nanoparticles in environmental media, plants, cells, and animals. As an example, Luo et al. incorporated the europium chelate Eu-β-diketonate into polystyrene (PS) particles, which had a diameter of 200 nm. They then used time-gated luminescence to visualize the PS-Eu particles in roots and shoots by analyzing the time-resolved fluorescence of the Eu chelate.¹⁴ Chen et al. labeled low-density lipoprotein (LDL) nanoparticles with the Cy7 (DiR) fluorophore and observed clear and intense fluorescence in the cytoplasm of KB cells (human epidermoid carcinoma cells) under a confocal microscope.¹⁵ Bindini et al. investigated the in vivo degradation of mesoporous silica nanoparticles (MSNs) by using radioactive labeling with ⁸⁹Zr, with the radiolabel localized either in the core or in the shell.¹⁶ Regarding sensor types, three labeling methods have been reported, including fluorescent, radiological, and metallic labeling. Among these methods, fluorescent labeling is the most commonly used to identify nanoparticles in media or cells, while radiological labeling is a convincing method for detecting nanoparticles in vivo.

In this Review, we systematically compare the advantages and limitations of fluorescent, radiological, and metallic labeling methods from the perspective of the testing model, physicochemical properties of nanoparticles, sensitivity, and other factors. Based on these comparisons, we propose a scheme for selecting appropriate labeling methods. Our review aims to provide comprehensive insights into the labeling techniques of nanoparticles, which will help in the development of effective methods for nanoparticle detection and tracking.

2. Fluorescent Labeling

Fluorescent labeling is a highly sensitive and selective technique that enables real-time monitoring and visualization of the behavior and distribution of nanoparticles in biological and environmental systems. This technique involves delicately modifying fluorescent molecules on the surface of nanoparticles or directly encapsulating them inside. It generally consists of four steps for fluorescent modification. This technique involves delicately modifying molecules on the surface of nanoparticles or directly encapsulating them inside. It generally consists of four steps. First, a suitable fluorescent molecule or dye that is compatible with both the nanoparticle surface and the imaging system being used is selected. Next, the nanoparticle surface is activated by using a suitable chemical or physical method, which may involve using a linker molecule or modifying the surface chemistry to introduce reactive groups that can react with the fluorescent molecule. After that, the fluorescent molecule or dye is attached to the nanoparticle surface by using a suitable coupling reaction, with reaction conditions depending on the specific chemistry being used. Finally, the fluorescently labeled nanoparticles are purified using size-exclusion chromatography, centrifugation, or other suitable methods to remove any unbound fluorescent molecules and characterized further by ultraviolet–visible (UV–Vis) spectroscopy, fluorescence spectroscopy, and microscopy. Achieving an efficient labeling reaction and ensuring the retention of the desired properties of nanoparticles require careful optimization of the process.

2.1. Covalent Labeling

Covalent conjugation is a common technique used for stable and specific labeling of nanoparticles on their surfaces for biological imaging and tracing. This method creates a stable bond between the fluorophore and the nanoparticle, preventing the fluorophore from dissociating or leaching off of the nanoparticle surface. Various methods for covalently conjugating fluorophores to nanoparticles exist, including silane coupling,^6,17 1-ethyl-3-(3-(dimethylamino)propyl) carbodiimide (EDC)/N-hydroxysuccinimide (NHS) coupling,¹⁸⁻²¹ and click chemistry.²²⁻²⁴

Silane coupling is the most commonly used method for functionalizing the surface of inorganic nanoparticles, including metal oxides, metals, metal sulfates, and MXenes. The surface of these nanoparticles contains sufficient hydroxy groups that can react with triethoxysilyl groups by a dehydration reaction. Active chemical groups such as amine, thiol, epoxy, and vinyl can be grafted onto particle surfaces by reacting with (3-aminopropyl)-triethoxysilane (APTES), 2-triethoxysilylethanethiol, (3-glycidoxypropyl)-triethoxysilane, and triethoxyvinylsilane, respectively (Figure 1A). The functionalized nanoparticles can then be labeled separately using various fluorophores, such as fluorescein isothiocyanate (FITC), tetramethylrhodamine-5-iodoacetamide (5-TMRIA), 7-amino-4-methylcoumarin (AMC), and indocyanine green-thiol (ICG-SH), via nucleophilic addition, nucleophilic substitution, epoxide ring opening, and click chemistry reactions, respectively.²⁵ For instance, Cai et al. used this method to label Fe₂O₃ rods, spheres, and plates by FITC to visualize their distribution in macrophages and epithelial cells by confocal microscopy.⁶ Interestingly, all Fe₂O₃ nanoparticles were internalized into lysosomes by endocytosis, indicating that the shapes of Fe₂O₃ had little impact on their subcellular distributions. In contrast, Liu et al. found a time-dependent distribution pattern of FITC-labeled Fe–N-doped graphene (FeNGR) nanosheets, including association with cytoplasmic membranes at 0–6 h, lysosomal internalization at 6–24 h, and release into the cytoplasm after 24 h²⁶ (Figure 1B). The cellular distributions may significantly impact the biological effects of nanoparticles. FeNGR associated with the plasma membrane could mimic the biological functions of NADPH oxidase to catalyze the generation of super oxides and boost immune activity, while FeNGR released into the cytoplasm may elevate the levels of NAD⁺ to speed up metabolic flux.²¹

Figure 1.

Covalent labeling to visualize nanoparticles in cells. (A) Schematic image showing the triethylsilane-based labeling process. (B) FeNGR was labeled by FITC conjugated with BSA and denoted as FITC-FeNGR (green). HepG2 cells treated with 10 μg/mL FeNGR for 6 and 24 h were stained with Hoechst 33342 (blue) and Alexa Fluor 594-conjugated wheat germ agglutinin (red) to visualize the nuclei and membrane in cells (scale bar = 10 μm). Reproduced from ref (26). Copyright 2023 American Chemical Society.

EDC/NHS coupling is an easy and effective method of connecting amine and carboxyl groups through a dehydration reaction in aqueous solutions. Carboxyl groups are found in many carbonaceous nanoparticles in the environment, including nanoplastics, carbon black, carbon nanotubes, and graphene oxides. Fluorophores containing amine groups, such as 2-aminoacridone, dansylcadaverine, and 7-amino-4-methylcoumarin, can be directly attached to nanoparticle surfaces. For example, graphene oxides (GOs)^18,19 and transition metal disulfates²⁰ (TMDs) containing sufficient carboxyl groups can be functionalized with FITC-labeled albumin for cellular imaging. Notably, even though both GO and TMDs are 2D nanosheets, they exhibit distinct distributions in THP-1 cells. GO primarily associates with the plasma membrane and causes membrane damage, while TMDs are internalized into lysosomes, resulting in lysosomal ferrous ion release, lipid peroxidation, and ferroptosis. Chen et al. used this labeling reaction to conjugate NH₂-PEG-Cy5.5 to the surface of trisodium salt (TETT silane)-modified MnO NPs. The strong tissue penetration of the 710 nm emissions from Cy5.5 enables near-infrared fluorescence (NIRF) imaging of MnO-PEG-Cy5.5 NPs in mouse brain tumors. As MnO has a strong magnetic resonance (MR) signal, the resulting MnO-PEG-Cy5.5 can act as a dual-modal (MR/NIRF) imaging nanoprobe, providing anatomical information from deep tissue inside the body and more sensitive information at the cellular level.²⁷

Click chemistry is a class of highly efficient reactions commonly used to connect two molecular moieties. It involves the use of chemoselective reactions that proceed under mild reaction conditions and typically result in high yields of the desired product. There are four types of click chemistry reactions: cycloaddition reactions (e.g., CuAAC,²⁸ SPAAC,²⁹ IEDDA³⁰), nucleophilic ring-opening reactions (e.g., using epoxides or aziridines), carbonyl chemical reactions of nonaldol reaction (e.g., the formation of oxime ethers or hydrazones), and addition reactions (e.g., thiol–ene or thiol–isocyanate) (Figure 2). Among them, the copper-catalyzed azide–alkyne cycloaddition (CuAAC) reaction is the most commonly used method to label nanoparticles. In this reaction, an azide-functionalized fluorophore reacts with an alkyne-functionalized nanoparticle in the presence of a copper catalyst, resulting in the formation of a covalent bond between the azide and alkyne. The azide groups can greatly affect the labeling reaction. Dirks et al. synthesized several 3-azido coumarin-terminated poly(ethylene glycol) (PEG) chains to investigate their conjugation with alkyne-functionalized bovine serum albumin (BSA). Interestingly, profluorescent 3-azido coumarin-terminated polymers can react with an alkyne-functionalized protein to produce a strongly fluorescent triazole-linked conjugate. Copper-catalyzed azide–alkyne cycloaddition (CuAAC) conjugation was shown to be very efficient and proceeded rapidly.³¹ Cao et al. reported a catalyst-free thiol–yne click reaction for the fabrication of aggregation-induced emission (AIE)-active fluorescent polymeric nanoparticles (FPNs) in a short reaction time. Specifically, an AIE-active dye with two alkynyl end groups (PhE-OE) was conjugated with thiol-group-containing polymers (PEGMA-IA-SH) via the catalyst-free thiol–yne click reaction to obtain AIE-active amphiphilic copolymers (PEGMA-PhE). Then, the resulting PEGMA-PhE copolymers could self-assemble into polymeric nanostructures, in which the hydrophobic PhE-OE was in the core while the hydrophilic components were on the surface. The AIE-active FPNs emitted intense fluorescence and had excellent dispersion in aqueous solution. They can be easily internalized by cells through a nonspecific route for cell imaging.³²

Figure 2.

Types of click reactions and representative examples. (A) Copper-catalyzed azide–alkyne cycloaddition. (B) Noncatalyzed azide–alkyne cycloaddition. (C) Nucleophilic ring-opening reaction of epoxy compounds. (D) Nucleophilic ring-opening reaction of aziridine compounds. (E) Carbonyl condensation reaction to form hydrazone. (F) Carbonyl condensation reaction to form oxime. (G) Thiol–ene addition reaction. (H) Thiol-isocyanato addition reaction.

2.2. Noncovalent Labeling

Noncovalent labeling is much simpler than covalent labeling and often involves a one-step reaction, but it may release bound fluorophores in complex environmental or biological media. This is a surface labeling method and may affect the surface chemistry of nanoparticles. There are four types of noncovalent labeling approaches: adsorption, electrostatic interactions, hydrophobic interactions, and host–guest interactions (Figure 3). Many nanoparticles have large surface areas that can adsorb small molecules or proteins for fluorescent labeling. For example, graphene oxide (GO) is a promising two-dimensional nanomaterial with a large surface area and abundant reactive surface groups, which make it a good adsorbent for various dyes in aqueous solutions. Liao et al. optimized the adsorption of rhodamine 123 (R123) onto GO by changing different culture media and the pH of the solutions. Under optimal conditions, they observed the cellular uptake and therapeutic release process of R123/GO using fluorescence microscopy.³³ Additionally, Kundu et al. modulated the adsorption of rhodamine 6G onto the GO surface by exploiting the hydrophobic interaction of Pluronic block copolymers with GO. In the presence of these pluronic aggregates, the quenching of fluorophores was minimized, and they observed an increase in the cellular uptake of GO using fluorescence lifetime imaging microscopy (FLIM).³⁴

Figure 3.

Noncovalent labeling approaches. Noncovalent labeling is a facile way to conjugate tags and light nanoparticles by adsorption, electrostatic interaction, host–guest interaction, and hydrophobic interaction.

Some nanoparticles, such as gold nanoparticles, have a surface charge that can interact with oppositely charged molecules through electrostatic interactions. Santra et al. embedded the positively charged dye tris (2,2′-bipyridine) ruthenium(II) (Ru(bpy)₃²⁺) into negatively charged silica particles by electrostatic adsorption to form fluorescent Ru(bpy)₃²⁺@SiO₂ NPs, which have been widely used in biological imaging and distribution.³⁵ Wang et al. adsorbed a negatively charged dye molecule (fluorescein sodium) on positively charged poly(methyl methacrylate) (PMMA) nanospheres and adsorbed a positively charged dye molecule (rhodamine B) on negatively charged PMMA nanospheres to synthesize two kinds of fluorescent dye@PMMA nanoparticles. These fluorescent nanospheres can easily dye blood cancer cells and track their distribution in vivo.³⁶

Hydrophobic interaction is another widely used strategy to label hydrophobic nanoparticles (e.g., nanoplastics and BN nanosheets) that can interact with hydrophobic fluorophores. These nanoparticles are often used as insulating materials in commercial products. High-sensitivity detection and tracking of these invisible materials are critical for their safety assessments. However, it is a significant challenge to visualize nanoinsulators in environments and biological systems because of their inert chemical surface, which prevents covalent functionalization. Zheng et al. found that fluorogenic molecules could be stably anchored on BN nanosheets by hydrophobic interactions and restrict intramolecular motions. Nanoscale insulators could be selectively lighted among 18 engineered nanomaterials by blockage of the nonradiative path and activation of the radiative electron transition of fluorogenic molecules. They found steric-restrictions-induced emission (SRIE). The SRIE-based lighting strategy enabled single-particle-level detection of nanoscale insulators in aqueous solutions and tracking of their distributions in complex biocontexts of cells and animal tissues.³⁷ This lighting strategy is a powerful technique to understand the fate of nanoinsulators in environments and biological systems and will greatly facilitate nanotoxicity assessment. Tian and co-workers synthesized a H₂O-dispersible NIR-II fluorescent nanoparticle containing a π-conjugated fluorescent molecule COi6-4Cl with an A–D–A configuration. Four bulky side chains of COi6-4Cl molecules induced restricted molecular vibrations and distinct intermolecular steric hindrance, suppressing the nonradiative decay in the dispersed and aggregated states. The fluorescent stabilization of COi6-4Cl nanoparticles makes them a distinguished material in NIR-II fluorescence-guided photodynamic therapy (PDT).³⁸

Certain nanoparticles can form host–guest complexes with small molecules or proteins. Host–guest interactions involve two molecules or materials that can form complexes through unique structural relationships and noncovalent binding. For example, Zhou and colleagues synthesized self-assembled supramolecular amphiphilic fluorescent nanoparticles (TP5/Zn/PM NPs) by using a host–guest complexation reaction. The host molecule was terpyridine-modified pillar arenes coordinated with Zn²⁺ (TP5/Zn), while the guest molecule was polyethylene glycol (PM). The visualization of the nanoparticles within living cells was achieved using a laser scanning confocal microscope.³⁹ Similarly, Song et al. developed NIR-II gold nanoclusters (Au NCs) coated with host cyclodextrin (CD) to label guest proteins with fluorescence. The stable macrocycle-based host–guest complexation of CD-Au NCs allowed efficient labeling of proteins/antibodies functionalized with guest molecules, thereby enabling the tracking of their physiological behavior in vivo.⁴⁰

2.3. Encapsulation of Fluorescent Quantum Dots by Core–Shell Structure

Some fluorophores could be encapsulated in the nanoparticle structure for stable tracing. Quantum dots (QDs) are nanocrystals made from semiconducting materials such as cadmium selenide or cadmium sulfide, typically with diameters ranging from 2 to 10 nm. At this scale, their properties differ significantly from those of their bulk counterparts, resulting in unique optical and electronic properties. When a fluorescent QD absorbs energy, an electron in the material is excited to a higher energy state. As the electron returns to its ground state, it emits light of a specific color or wavelength, which can be adjusted by varying the size and composition of the QD. Fluorescent QDs have several advantages over traditional fluorescent dyes or proteins: they have high brightness, broad excitation spectra, and narrow emission spectra and are highly photostable, meaning that they can be excited and emit light repeatedly without degrading or losing their brightness.⁴¹ Encapsulating fluorescent QD cores into nanoparticles creates a core–shell structure to enhance the optical properties of the nanoparticles. This approach is commonly used to improve the stability, biocompatibility, and functionality of fluorescent nanoparticles.⁴² The shell can be made of various materials such as polymers, phospholipids, silica, or metallic compounds. Polymers such as polyethylene glycol (PEG), poly(vinyl alcohol) (PVA), poly(lactic-co-glycolic acid) (PLGA), and polystyrene (PS) have been used extensively to form shell layers via self-assembly, polymerization, emulsion, and nanoprecipitation.

Phospholipids are amphiphilic molecules with both hydrophobic and hydrophilic properties, making them ideal for forming shells around fluorescent QDs. The phospholipid layer, such as a micelle or liposome, also provides a biocompatible surface for the particles, making them suitable for use in biological systems. The encapsulation process can be achieved through various methods, including thin-film hydration, reverse-phase evaporation, and microemulsion techniques.⁴³⁻⁴⁵ Howes et al. synthesized magnetic-fluorescent semiconductor polymer nanospheres (MF-SPNs) by coencapsulating hydrophobic conjugated polymers and iron oxide nanoparticles in phospholipid micelles. They successfully demonstrated the uptake behavior of MF-SPNs by SH-SY5Y neuroblastoma cells through bright and stable fluorescence.⁴⁶ Chen et al. encapsulated fluorescent QDs into liposomes and provided a methodology to quantify the number of nanoparticles inside each liposome by fluorescence correlation spectroscopy.⁴⁷ Additionally, Tansi et al. reported a new fluorescence-activatable, PEGylated liposome formulation containing high concentrations of the NIR dye DY-676-COOH, which has high potential for in vivo optical imaging.⁴⁸

Given the transparent feature of SiO₂ to visible light, core–shell labeling is an ideal method to label silica nanoparticles. Silica encapsulation can be achieved through several methods, including Stöber synthesis, reverse microemulsion, and sol–gel methods. Zane et al. synthesized fluorescent silica nanoparticles labeled with CdSe/CdS/ZnS QDs by the Stöber synthesis method and examined their presence in various organs by combining confocal fluorescence microscopy with inductively coupled plasma–mass spectrometry.⁴⁹ Furthermore, the environmental stability of silica makes it an excellent surrogate particle for mass tracking under harsh conditions such as geothermal plants, nuclear reactors, industrial processes, and oil fields. Hubbard et al. employed Stöber methods to synthesize silica nanoparticles labeled with ZnS/Mn QDs and tested their tolerance over a pH range of 3–13. These tracer particles are promising for mass production, tolerating harsh environments, and enduring debris.⁵⁰

2.4. FRET Effect on Light Nanoparticles

Förster resonance energy transfer (FRET) refers to the transfer of energy from a donor fluorophore to an acceptor fluorophore through dipole–dipole interaction. The efficiency of energy transfer in a FRET system depends strongly on the distance between donor and acceptor molecules and requires a good overlap between donor emission and acceptor absorption bands. As the distance between the two fluorophores increases, the efficiency of energy transfer decreases, leading to an increase in the fluorescence from the donor fluorophore. Because of this fact, the FRET effect has been widely researched in fluorescent imaging. Ma et al. constructed a target-triggered fluorescent nanoprobe composed of a ratiometric pH-sensitive dye (N-carboxyhexyl derivative of 3-amino-1,2,4-triazole fused with 1,8-naphthalimide, ANNA), a near-infrared dye (Cy5.5), and biocompatible Fe₃O₄ nanoparticles. ANNA was linked through a peptide substrate of matrix metalloprotease-9 (MMP-9) with Fe₃O₄ nanoparticles to establish a FRET system for sensing the pH of the tumor microenvironment. Once the peptide linker connecting ANNA to the Fe₃O₄ nanoparticles is cleaved by MMP-9 in tumors, it will induce the activation of the pH-sensing fluorophore on Fe₃O₄. This dual-ratiometric target-triggered fluorescent probe could be exploited for the simultaneous quantitative visualization of protease activity and pH in the tumor microenvironment.⁵¹ Cai et al. developed a rapid fluorescence “switch-on” assay to detect trace amounts of GSH based on carbon dot-MnO₂ nanocomposites. The formation of carbon dot-MnO₂ nanocomposites quenches the fluorescence of carbon dots efficiently through FRET. However, the presence of GSH reduces MnO₂ nanosheets to Mn²⁺ ions peeling off from carbon dots, resulting in the recovery of the fluorescent signal.⁵² Moreover, Wu et al. prepared bioconjugates of PEI–PLNPs with Ab-AuNPs based on the FRET effect, providing a sensitive and specific persistent luminescence probe for the detection of AFP in biological fluids and for imaging AFP excreted during cancer cell growth.⁵³ Since this labeling method requires a sophisticated modification of nanoparticle surface, the functional groups may affect the interactions between nanoparticles and biological systems.

3. Radiological Labeling

Radioisotope tracing technology utilizes radioisotopes as tracers to investigate the dynamics of previously obscure processes. By monitoring the movements and changes of tracers, researchers can gain insight into the behavior and fate of nanomaterials in organisms. Although radioisotope tracing involves exposing organisms to radioactivity, it is still widely considered the best method for studying the various processes involved in the absorption, distribution, metabolism, excretion, and environmental behavior of nanomaterials in organisms.

3.1. Chelator-Based Radiolabeling

The use of radioisotope-tracing technology is widespread in labeling nanodrugs for diagnosis and treatment. Nonmetallic radionuclides, such as ¹⁸F and ¹³¹I, are labeled through direct covalent bond formation with the compounds. However, metallic radionuclides, such as⁶⁴Cu and ⁸⁹Zr, often require coordination chemistry approaches and the use of chelators to attach them effectively to the desired nanoparticle.⁵⁴⁻⁵⁶

The superparamagnetic properties of γ-Fe₂O₃ and Fe₃O₄ allow for their use in drug delivery systems and tracking distribution in real time, including targeting specific lesion sites. In addition to serving as clinical magnetic resonance imaging contrast agents, these nanoparticles can be radiolabeled using chelator-based methods.⁵⁷⁻⁵⁹ Chelator-based methods using DOTA, NOTA, and bis(dithiocarbamate) bisphosphonate have been reported for radiolabeling ⁶⁴Cu on iron-oxide nanoparticles,⁶⁰⁻⁶² while poly(acrylic acid) , bisphosphonate derivatives, and DMSA are used to radiolabel ⁹⁹Tc (Table 1).⁶³⁻⁶⁵ Sun et al. developed a fast, facile surface coating procedure to apply one-step ¹⁸F radiolabeling of iron oxide nanoparticles for in vivo dual PET/MR imaging using ¹⁸F-AlF/NOTA.⁶⁶

Table 1. Radiological Labeling of Representative Nanoparticles.

NP type	radiolabeling method	radionuclide	ref
iron-oxide NP	chelator-based	64Cu	(105)
			(60)
			(62)
		99mTc	(64)
			(63)
			(65)
	chemical adsorption	18F	(66)
		64Cu	(106)
		89Zr	(105)
			(107)
		223Ra	(108)
SiO2	chelator-based	117Lu	(68)
		22Na	(69)
	chemical adsorption	89Zr	(99)
		68Ga	(99)
		111In	(99)
		90Y	(99)
		177Lu	(99)
		64Cu	(99)
		45Ti	(101)
liposome	chelator-based	111In	(72, 74)
		64Cu	(73)
		99mTc	(74)
	directly labeling	99mTc	(109, 110)
Au NP	chelator-based	99mTc	(76−78)
	hot-plus-cold precursors	195Au	(79)
		199Au	(97)
		198Au	(98)
nanoplastic	hot-plus-cold precursors	14C	(80, 81, 85)
graphene	hot-plus-cold precursors	14C	(88−90)
	directly labeling	64Cu	(102)
		68Ga	(103)
CuS NP	hot-plus-cold precursors	64Cu	(82, 111)
RBC-EM	directly labeling	99mTc	(112, 113)

Silica nanoparticles are solid materials that have a honeycomb-like porous structure with numerous mesopores, which can absorb or encapsulate relatively large amounts of bioactive molecules.⁶⁷ Although mesoporous silica nanoparticles lack optical or magnetic properties, they have adjustable textural features, such as porosity, crystallinity, and morphology, that make them useful in studies. For instance, Zhang et al. labeled ¹⁷⁷Lu to an ultrasmall fluorescent core–shell silica nanoparticle (Cornell prime dots or C’dots) surface measuring approximately 6 nm in diameter to develop a new radiotherapy for melanoma, with mice showing a longer survival time.⁶⁸ Additionally, Al Faraj et al. surface labeled silica nanoparticles with ²²Na using 4-amino-benzo-15-crown-5 and developed a radiolabeling method suitable for departments that lack onsite generators or cyclotrons.⁶⁹

Liposomes are spherical vesicular organic nanoparticles composed of lipid bilayers that have been extensively studied for diagnostic and therapeutic purposes, particularly in the delivery of anticancer drugs.⁷⁰ Similar to other nanoparticles, the pharmacokinetics of liposomes in living organisms depend on their size, surface charge, and delivery pathways.⁷¹ Radioactive labeling of liposomes dates back to the 1980s, when they were loaded with ¹¹¹In for imaging purposes.⁷² Henriksen et al. loaded the chelator DOTA with ⁶⁴Cu²⁺ ions into liposomes, which caused spontaneous diffusion of copper across the lipid bilayer and subsequent trapping.⁷³ Helbok et al. evaluated DTPA-derivatized lipid-based nanoparticles (DTPA-NP) radiolabeled with different radiometals, including ¹¹¹In and ^99mTc.⁷⁴

Gold nanoparticles have exceptional catalytic activity and optical properties, making them invaluable in a variety of fields such as nanodevice manufacturing, nanobiotechnology, nanobiomedicine, and nanopharmacology. Their outstanding physicochemical, optical, and photoacoustic contrast properties, combined with high biocompatibility, make them prime candidates for various applications in nanomedicine.⁷⁵ Several chelator-based labeled gold nanoparticles exist, such as those labeled with ^99mTc with DTPA, DOTA, HYNIC-TOC, and others.⁷⁶⁻⁷⁸

However, given that chelators are immobilized on particle surface, endogenous proteins can cause transchelation or detachment of molecular chelators that strip gold nanoparticles of their radiolabels, resulting in inaccurate biodistribution images.⁷⁹ Additionally, chelators may negatively impact the biological behavior of radio-immunoconjugates, and selecting the appropriate chelator for a specific isotope can be challenging or impossible.^80,81 Furthermore, concerns about complex coordination chemistry, altered pharmacokinetics of carriers, and the potential detachment of radioisotopes during imaging have emphasized the need for simpler and better radiolabeling techniques for future applications.

3.2. Nonchelator Radiolabeling

Radioactively labeled nanomaterials prepared by this method have almost the same physical and chemical properties as stable nanomaterials except for their radioactivity. This means that nanoparticles synthesized by this method could be used in conventional nanotoxicity or environmental pollution research.

3.2.1. De Novo Synthesis

This strategy, exclusive to inorganic nanomaterials, often involves adding trace amounts of radioactive (or “hot”) precursors together with nonradioactive (or “cold”) precursors into the backbone of nanoparticles. It is a simple and direct method and is also the most widely used method for acquiring radioactive nanoparticles, as the nuclides are attached on the backbone of nanoparticles. With this method, the selected radioactive isotopes can be embedded well into the lattice of the final nanocrystal, resulting in intrinsically stable radioactive nanoparticles.⁸²

Nanoplastics, which refer to plastic debris smaller than 1 μm, are considered to be the most hazardous component of marine plastic litter.^83,84 As a nanoparticle pollutant, it is necessary to maintain its original physical and chemical properties during labeling to trace its migration, transformation, and fate in complex environmental systems. Therefore, intrinsic radiolabeling is the preferred method for nanoplastic labeling. Recently, a novel method for synthesizing ¹⁴C-labeled nanoplastics was developed to enhance the efficiency of investigating their in vivo behavior in bivalve mollusks by Al-Sid-Cheikh et al.⁸⁵ The method offers several benefits compared to earlier proposed techniques,^80,81 including the use of less than 10 mg of radioactive styrene monomer, which overcomes the challenge of high cost. Furthermore, this method produces ¹⁴C-labeled nanoplastics with specific activity levels that are suitable for research at environmentally relevant concentrations in the parts per billion range.⁸⁶

Graphene is a lattice-stacked layer of carbon atoms that produces a single layer of graphite. These nanomaterials have attractive optical and mechanical properties and have been used for drug delivery and biomedical imaging.⁸⁷ Because graphene is a full-carbon material, the addition of any other element will affect the unique position of carbon within it. Therefore, C-14 appears to be the best choice for radiolabeling graphene and has been widely used to track graphene materials in diverse organisms⁸⁸ (Figure 4). Guo et al. synthesized ¹⁴C-labeled graphene by following these steps: first, FeCl₂·4H₂O and NH₄H₂PO₄ were added to ultrapure water at 50 °C. Then, an ethanol solution containing dodecylamine was added to the mixture. Afterward, ¹⁴C-labeled phenol, dissolved in ethanol solution, was added to the dry powder, and the mixture was dried at 40 °C for 12 h. The mixture was then heated at 700 °C for 12 h under argon. To synthesize ¹⁴C-labeled graphene, the obtained black powder and 40 mL of 37% hydrochloric acid must be transferred into a Teflon-lined autoclave and sealed before being heated in an oven at 180 °C for 24 h.⁸⁹ As reported in this study, ¹⁴C-graphene exhibits the same chemical and physical properties as graphene and provides accurate quantification for in vivo or environmental applications.⁹⁰ Lu et al. found that ¹⁴C-labeled graphene was modified in enzyme-mediated reactions, and the results in Daphnia magna revealed a substantial decrease in the bioaccumulation and toxicity of graphene.⁹¹ By detecting radiolabeled CO₂, Dong et al. found that the graphene in soil was degraded by a heterogeneous Fenton reaction into the gas phase with H₂O₂ and ferric oxide that occurred naturally in red soil.⁹² With the support of radioactive tracing, Su and co-workers studied the settling behavior of graphene in different water system with extremely low concentration (158 μg/L), and green algae was found to cospittle with graphene and consequently alter the graphene phase distribution.^93,94 Lu et al. revealed that ¹⁴C-labeled graphene can cross the intestinal wall and enter intestinal epithelial cells and the bloodstream in zebrafish, benefiting from the advantage of ¹⁴C labeling, and the internal exposure dose of zebrafish intestinal epithelial cells was accurately quantified.⁹⁰ In another study, Lu et al. exposed two sizes of ¹⁴C-labeled graphene (smaller lateral dimensions (SLG, 20–40 nm) and larger lateral dimensions (LLG, 330–630 nm)) into mice by intravenous injection and indicated that ¹⁴C-labeled graphene mainly accumulated in the liver: approximately 60.91 and 74.13% of the total liver uptake for SLG and LLG in liver.⁸⁸ Dong et al. observed that graphene has the ability to enter the root cells of wheat through the apex of root hair and move through the symplastic pathway to the vascular bundle for further transport to the shoot.⁹⁵ Furthermore, Dong et al. investigated the fate of ¹⁴C-labeled graphene in aquatic food webs and found that trophic transfer may lead to the bioaccumulation of graphene in organisms. To assess the bioaccumulation and biomagnification of graphene, the body burden factor (BBF) and trophic transfer factor (TTF) were analyzed for each organism and food chain. The calculated TTFs of 0.2–8.6 in the food chain from Escherichia coli to Tetrahymena thermophila suggest substantial trophic transfer potential in this aquatic ecosystem. These findings imply that graphene, if introduced into this ecosystem, has the potential to biomagnify, with higher trophic organisms such as T. thermophila being at a greater risk of accumulating significant amounts of graphene. This poses potential risks to the ecosystem structure and stability over time, underscoring the need for further research to evaluate the ecological risks associated with graphene and develop strategies for mitigating its environmental impact.⁹⁶ These findings provide important insight into the potential ecological risks associated with graphene exposure and validate the versatility and precision of ¹⁴C de novo synthesis labeling and tracing methodologies in multiple settings.

Figure 4.

Uptake, transfer, and environmental fate of graphene clarified by labeling with ¹⁴C.

The radioactive nuclide ⁶⁴Cu is widely used, and copper sulfide nanoparticles are ideal intrinsic radiolabel nanoparticles due to their stability and biocompatibility. ⁶⁴Cu-labeled copper sulfide (CuS) could be prepared using a hot-plus-cold precursor method by adding ⁶⁴CuCl₂ (hot precursor) to a mixture of CuCl₂ and Na₂S (cold precursors) in a 95 °C water bath.⁸²

Various analytical methods are available to estimate gold in different organs, but radiolabeling gold with radioactivity is a highly effective and accurate method for quantifying and understanding the pharmacokinetics of AuNPs. It has been found that the scintillation scanning radio-counting method has greater sensitivity than any other analytical technique. Kreyling et al. labeled 20 nm AuNPs with ¹⁹⁵Au through aerosol spark ignition and found that inhalation of these nanoparticles through intratracheal administration caused deposition, preferably in the caudal lungs regardless of the age of the rat. It is noteworthy that AuNPs exhibited significant retention in soft tissue, including muscle, connective tissue, fat, and skin, which has not been reported in several other studies.⁷⁹ Zhao et al. synthesized 5 and 18 nm gold nanoparticles labeled with ¹⁹⁹Au using seed-mediated growth and hot-plus-cold precursors. The ¹⁹⁹Au-AuNPs demonstrated an improvement in the biodistribution profile when conjugated with DAPTA for CCR5 targeting, and they also showed a significant enhancement in tumor SPECT/CT imaging sensitivity and specificity.⁹⁷ In another study, ¹⁹⁸Au-labeled AuNPs were synthesized using the H¹⁹⁸AuCl₄ precursor.⁹⁸ It is worth noting that ¹⁹⁸Au exhibits the Cerenkov luminescence phenomenon and has potential as a Cerenkov luminescence imaging agent.

3.2.2. Chemical Adsorption

One strategy for radiolabeling nanomaterials involves utilizing their strong adsorption ability to adsorb radioactive ions or colloids from solutions onto their surfaces. This method is straightforward, but the labeled radioactive nuclides may detach in complex body environments, requiring further verification of the labeled substance’s stability. Silica nanoparticles can be labeled with radionuclides using chelator-based radiolabeling methods, although chemical adsorption is often used due to its effectiveness. The stability of radiolabeled silica nanoparticles depends on the hardness of the isotope and its binding affinity. In a study by Shaffer et al., six different isotopes (⁸⁹Zr, ⁶⁸Ga, ¹¹¹In, ⁹⁰Y, ¹⁷⁷Lu, ⁶⁴Cu) were labeled on silica nanoparticles and tested for serum stability by incubating them in 50% FBS at 37 °C.⁹⁹ Among them, copper was the least oxophilic and had weaker retention owing to the affinity each isotope has for the oxygen-rich matrix.¹⁰⁰ Chen et al. synthesized uniform MSN with an average particle diameter of approximately 150 nm using the affinity between strong Lewis acids and strong Lewis bases for labeling the particles with ⁸⁹Zr and ⁴⁵Ti.¹⁰¹ However, when using chemical adsorption methods, consideration should be given to whether changes in particle surface properties will affect their properties.

Graphene nanosheets possess plentiful π bonds on their surface, enabling them to be labeled with ⁶⁴Cu by simply mixing reduced graphene oxide (RGO) and graphene oxide (GO) with ⁶⁴Cu in 0.1 m of sodium acetate buffer. This was demonstrated by a study by Shi et al.¹⁰² Additionally, Sarpaki et al. described the radiolabeling of new functional graphene oxide composites using gallium-68. GO nanocomposites were directly radiolabeled with aqueous ⁶⁸GaCl₃ in just an hour.¹⁰³ However, the labeling efficiency and stability of the incorporated isotope are mostly reliant on the number of π bonds, making it challenging to guarantee the repeatability of the labeling efficiency.

3.2.3. Covalent Conjugation

The active surface of the nanoparticles provides an opportunity to react with ambient oxygen to form hydroxyl groups. Therefore, most nanoparticles could be functionalized with silane to acquire desired functionalities on their surface, such as NH₂, COOH, and phenyl groups (Table 1). Recently, Liu et al. developed a common approach for radiological labeling of nanoparticles by iodonucleoid. As shown in Figure 5A, hydroxyl groups on the nanoparticle surface could react with triethoxy (4-methoxyphenyl) silane (TEMOPS) in anhydrous dimethylformamide (DMF). After the dehydration reaction, methoxyphenyl groups were covalently attached to the particle surface. The resulting materials were dried and subjected to ¹²⁵I radiolabeling by an electrophilic substitution reaction. After purification by centrifugation to remove free ¹²⁵I, the labeled nanoparticles could be exploited to visualize the biodistribution of intravenously injected FeNGR²⁶ (Figure 5B) or the retention of microspheres in mouse tumors¹⁰⁴ (Figure 5C). Interestingly, FeNGR rapidly accumulates in the mouse liver in 6 h and is slowly eliminated in 9 days, while intratumorally injected microspheres are stably present in the tumor xenograft for at least 2 weeks. The ligands have a great impact on the stability of covalently labeled nanoparticles. Since there are no enzymes to digest silicon compounds in mammals, oxosilane ligands are highly recommended for surface functionalization.

Figure 5.

Radiological labeling of nanoparticles. (A) Schematic of radiological labeling of nanoparticles by ¹²⁵I. (B) SPECT/CT imaging of intravenous injected ¹²⁵I-FeNGR in mice. Reproduced from ref (26). Copyright 2023 American Chemical Society. (C) SPECT/CT visualizes ¹²⁵I-labeled microspheres in tumors. The dashed yellow circle indicates the tumor outline. Reproduced from ref (104). Copyright 2022 Elsevier Ltd.

4. Metal Labeling

Metal labeling is a widely used internal labeling technique to investigate the distribution and fate of nonmetallic nanoparticles (e.g., nanoplastics, polymers, or silica) in different environmental mediums or organisms. The process of metal labeling involves attaching or doping metal ions or nanoparticles to target nanoparticles. Metal indexes can be attached using various methods, including chelation, core–shell construction, or doping. The labeled nanoparticles can be detected using various analytical techniques, such as MRI, transmission electron microscopy (TEM), or inductively coupled plasma optical emission spectrometry (ICP-OES). Compared to fluorescent and radiological labeling, metal labeling allows for the selection of diverse metal elements as labeling indexes.

Lanthanide chelates have long luminescence lifetimes, large Stokes shifts, and sharp emission profiles that make them suitable for imaging nanoparticles in complex biological systems. Luo et al. used polystyrene (PS) nanoparticles doped with a europium chelate, Eu-β-diketonate (PS-Eu), to quantify the uptake of PS-Eu nanoparticles in wheat (Triticum aestivum) and lettuce (Lactuca sativa) using inductively coupled plasma–mass spectrometry (ICP-MS). They found that PS-Eu particles mainly accumulated in the roots and a few of them were transported to the shoots. They also visualized PS-Eu particles in the roots and shoots through the time-resolved fluorescence of the Eu chelate.¹⁴ Li et al. synthesized an amphiphilic tris(dibenzoylmethanato) europium(III) coordinated copolymer that exhibited good biocompatibility and emitted strong red luminescence. The copolymer could self-assemble into micelles and could be easily tracked in A549 tumor cells and zebrafish larvae.¹¹⁴ In addition, metal-coordinated polymer gels have been considered a promising strategy for hybrid nanoplatforms used for imaging and therapeutic applications due to their noncrystalline and elastic nature. Lim et al. reported functional gadolinium-coordinated nanogels (GdNGs) for in vivo tumor imaging with long blood circulation. The Gd³⁺-rich structure of GdNGs could be visualized by MRI with an enhanced T2 scan.¹¹⁵

As surface metal tags may influence the behavior of nanoparticles and dissociate from their surface in environmental or biological matrices, metal doping or incorporating metal cores inside nanoparticles presents an advantage in stably associating with the nanoparticles without affecting their surface chemistry. This technique is especially useful for detecting nanoplastics and SiO₂, two substances that are widely spread in the environment. Different metal cores, such as Au,¹¹⁶ Ag,¹¹⁷ TiO₂,¹¹⁸ Fe_xO_y,¹¹⁹ and ZnO,¹²⁰ can be embedded into silica nanoparticles. Vitorge et al. embedded Ag nanoparticles in silica shells to investigate their transport, deposition, and release in a natural sand representative of an aquifer. The detection limit reached 2 × 10^–9 (mol/L).¹¹⁷ Mitrano et al. synthesized nanoplastics with a chemically entrapped metallic fingerprint (Pd), which can greatly aid studies to understand the underlying mechanisms, processes, and principles of nanoplastic behavior. They demonstrated the utility of this approach in simulating the activated sludge process of wastewater treatment to better understand the fate of nanoplastics in urban environments.¹²¹

In recent years, the effects of nanoplastics on aquatic biota have raised particular concerns due to their small size, which enables them to be taken up by cells and affect biota on a cellular level. Aquatic organisms can actively ingest nanoplastics or adsorb them to their surfaces and transfer them to higher trophic levels.¹²¹ To assess the uptake and chronic effects of nanoplastics on the survival and growth of the freshwater benthic macroinvertebrate Gammarus pulex, Koelmans and co-workers used metal-doped nanoplastics for standardized single-species sediment toxicity tests. They quantified the nanoplastic concentrations based on Pd concentrations measured by ICP–MS on digests and fecal pellets of the exposed organisms.¹²²

5. Multimodal Labeling

Significant progress has been made in these labeling methodologies, each of which exhibits a certain advantages. By combining the strengths of multiple labeling methods, researchers can obtain a more complete understanding of nanoparticle behaviors and biodistribution in interactions with biological systems. For example, Shibu et al. reported the preparation of photouncaging nanoparticles by combining photoluminescent ligands, CdSe/ZnS quantum dots, and super paramagnetic iron oxide nanoparticles.¹²³ They used MRI and NIR fluorescence imaging to examine the biological fate of the nanoparticles in mice. Furthermore, multimodal labeled NPs are promising candidates for biomedical applications. Harmsen et al. developed a dual-modal PET/NIRF nanoparticle for nongenomic labeling of human chimeric antigen receptor (CAR) T cells.¹²⁴ They used a modified Stöber method to synthesize silica nanoparticles containing a silane-appended near-infrared fluorophore and ⁸⁹Zr nuclide. The nanoparticles were coated with protamine and heparin to label of CAR T cells. Long-term whole-body CAR T-cell tracking was enabled by PET and NIRF in the ovarian peritoneal carcinomatosis model. The multimodal imaging may be extremely useful to examine the behavior of nanocomposites, such as the dissociation and transformation of protein corona on nanoparticles.¹³

6. Conclusions

Overall, fluorescent labeling is a highly sensitive, specific, and nondestructive technique, allowing multiplex-tag labeling, versatile uses, and real-time monitoring, but it has limitations in long-term and deep tissue imaging. Radiological labeling is the best choice for the detection of nanoparticles in deep tissues of large animals, such as mice, rats, rabbits and pigs, and allows dynamic imaging in vivo. However, radiology imaging merely allows three channels at X, γ, or β ray, and it is important to weigh the potential risks before using this technique. The risks can be mitigated by following appropriate safety protocols and using alternative imaging techniques when possible. Metal labeling is often a good quantitative method for the sensitive detection of labeled nanoparticles, but is rarely used for real-time monitoring (except for MRI contrast agents) and may cause potential toxicity. In regard to labeling a specific nanoparticle, research selecting an appropriate labeling method for nanoparticles requires careful consideration of multiple factors, including the type of nanoparticle, the purpose of labeling, the detection method, compatibility with the nanoparticles, safety considerations, and cost and availability.

• Type of nanoparticle: Different nanoparticles may require different labeling techniques depending on their size, shape, surface chemistry, and other physical and chemical properties. For example, metallic nanoparticles may be labeled using fluorescence, whereas organic nanoparticles could be labeled by metal tags.

• Purpose of labeling: The purpose of labeling will also influence the choice of labeling method. For example, if the goal is to track the nanoparticles in vivo, then deep penetration labeling techniques such as radiological or magnetic resonance imaging (MRI) may be preferred over fluorescence labeling.

• Compatibility with the nanoparticles: The labeling method should be compatible with the nanoparticles and not affect their physicochemical properties, stability, or biological activity. Some surface labeling methods should be carefully examined to determine whether the labeling tags significantly alter the agglomeration state, surface charge or hydrophilicity of nanoparticles, resulting in different behaviors in biological systems.

• Safety considerations: The safety of the labeling method and the labeled nanoparticles should also be considered. For example, radioactive labeling may pose safety risks and require special handling and disposal procedures.

• Cost and availability: The cost and availability of the labeling method and reagents should also be taken into account. Some labeling methods may be expensive or require specialized equipment, whereas others may be more accessible and cost-effective.

• Labeling efficiency: Although a high dense of labeling tags may greatly facilitate a sensitive detection of nanoparticles, extensive surface modifications may affect the surface chemistry of nanoparticles as well as their interactions with biological systems. To achieve sensitive detection and minimize the impacts of surface modification, the labeling efficiency are recommended to maintain at mass percentages of 0.1–10‰, 0.1–1%, and 0.5–10‰ for fluorescent, radiological, and metal labeling, respectively.

Author Contributions

^# H.C. and Q.H.: Equal contribution.

The authors declare no competing financial interest.