|Home | About | Journals | Submit | Contact Us | Français|
Theranostic nanomedicine is emerging as a promising therapeutic paradigm. It takes advantage of the high capacity of nanoplatforms to ferry cargo and loads onto them both imaging and therapeutic functions. The resulting nanosystems, capable of diagnosis, drug delivery and monitoring of therapeutic response, are expected to play a significant role in the dawning era of personalized medicine, and much research effort has been devoted toward that goal. A convenience in constructing such function-integrated agents is that many nanoplatforms are already, themselves, imaging agents. Their well developed surface chemistry makes it easy to load them with pharmaceutics and promote them to be theranostic nanosystems. Iron oxide nanoparticles, quantum dots, carbon nanotubes, gold nanoparticles and silica nanoparticles, have been previously well investigated in the imaging setting and are candidate nanoplatforms for building up nanoparticle-based theranostics. In the current article, we will outline the progress along this line, organized by the category of the core materials. We will focus on construction strategies and will discuss the challenges and opportunities associated with this emerging technology.
The term “theranostics” was coined to define ongoing efforts in clinics to develop more specific, individualized therapies for various diseases, and to combine diagnostic and therapeutic capabilities into a single agent. The rationale arose from the fact that diseases, such as cancers, are immensely heterogeneous, and all existing treatments are effective for only limited patient subpopulations and at selective stages of disease development. The hope was that a close marriage of diagnosis and therapeutics could provide therapeutic protocols that are more specific to individuals and, therefore, more likely to offer improved prognoses.
The emergence of nanotechnology has offered an opportunity to draw diagnosis and therapy closer. Nanoparticle (NP)-based imaging and therapy have been investigated separately, and understanding of them has now evolved to a point enabling the birth of NP-based theranostics, which can be defined as nanoplatforms that can co-deliver therapeutic and imaging functions. This is in a way an extension of the traditional theranostics but focusing more on “co-delivery”. It adds to the previous paradigm for allowing imaging to be performed not only before or after, but also during a treatment regimen. It is convenient that many nanomaterials are already imaging agents and can be readily “upgraded” to theranostic agents by mounting therapeutic functions on them. One underlying driving force of such a combination is that imaging and therapy both require sufficient accumulation of agents in diseased areas. This common targeting requirement brings the two research domains closer and, ultimately, will blur the boundary between them, since many techniques to enhance imaging can, at least in theory, be readily transferred to the therapeutic domain, and vice versa.
Targeting strategies can be varied immensely to suit the desired targets. In the case of cancer, it is a common approach to identify a biomarker that is aberrantly expressed on the surface of cancer cells, and then to load its cognate binding vector onto probes/carriers to achieve recognition and tumor homing. For nanoplatforms, the unique size scale of the particles enables achievement of an enhanced-permeability-and-retention (EPR) effect in tumor targeting. In all efforts, however, care has to be taken with the particles’ surfaces to avoid innate immunosystem recognition and to secure sufficiently long circulation half lives for the agents to reach their targets.
Nanoparticle-based imaging and therapy are each struggling to advance into clinical trials and, as descendants of the two, nanoparticle-based theranostics are still in their early stages of development. However, the push provided by advances in nanotechnology and the call for personalized medicine have already made nanoparticle-based theranostics a research hotspot. This review attempts to give a summary of the efforts made so far along this line. We will introduce theranostic agents, arranged by the category of their core nanomaterial, that hold potential in the theranostic setting. The techniques used to form linkage between nanoplatforms and functionally entities have been well developed and are summarized in Table 1. As mentioned above, most of the nanoplatforms to be described here already perform imaging functions and have been widely investigated for imaging related applications. However, imaging alone, without therapy, will not be the focus of this article, and readers are referred to several excellent reviews on that topic. Instead, we will focus on the build-up and application of theranostic agents, as well as the associated surface coating and coupling chemistry that may affect transport, delivery and release of cargos.
Iron oxide nanoparticles (IONPs) are nanocrystals made from magnetite or hematite. Despite spin surface disorders and spin canting effect, IONPs typically possess substantial saturation magnetization (Ms) values at room temperature, especially for those made from pyrolysis protocols with good crystallinity. Unlike the bulk materials, IONPs less than 20 nm are superparamagnetic--a state where particles show zero magnetism in the absence of an external magnetic field, but can become magnetized when there is one. The underlying mechanism is that at such small scale, the thermal energy is sufficient to overcome the anisotropy energy of each small magnet (nanoparticle), and this leads to random fluctuation of the magnetizations that, macroscopically, result in zero net coercivity and magnetic moment.
The superior magnetic properties of IONPs, along with their inherent biocompatibility and inexpensiveness, have made IONPs a material of choice in many bioapplications, such as contrast probes for magnetic resonance imaging (MRI). The high magnetic moments of IONPs make them effective in reducing T2 relaxation time, leading to signal attenuation on a T2 or T2* weighted map. When the particles are engineered with targeting specificity, such signal alterations can be harnessed to report abnormal biological activity.
The synthesis of IONPs has been well documented. Traditionally, IONPs are made in aqueous solution by co-precipitating Fe(II) and Fe(III) precursors. In order to confer colloidal suspendability to the particles, additives, typically hydrophilic polymers, are added during the particle formation process, which passivate the nanocrystal surface and protect against particle aggregation. A number of ligands including polyvinylpyrrolidone (PVP), dendrimer, polyaniline and dextran have been utilized for such purposes, with dextran and its derivatives being the most studied. As a matter of fact, several dextran-IONP formulas have entered or already passed clinical trials as MRI contrast agents. For instance, Feridex particles (AMAG Pharmaceuticals) are FDA approved for the detection of liver and spleen lesions, and their analog Combidex has entered into phase III clinical trial for lymph node imaging. In addition to helping improve colloidal stability of the particles, the polymer coating offers multiple chemical groups that are essential for the coupling of functional species. For instance, we have prepared polyaspartic acid (PASP) coated IONPs and coupled onto the particles RGD and 64Cu-DOTA for positron emission tomography (PET)/MRI dual imaging. Lee et al. have converted dextran-IONPs aminated with epichlorohydrin and ammonia, after which the particles can be readily conjugated with a wide range of biospecies.
Recently, high temperature decomposition has emerged as a useful strategy in nanoparticle preparation. Unlike traditional methods, where aqueous solution is used as the reaction medium, such pyrolysis synthesis takes place in organic solvent upon high temperature treatment. Due to the existence of highly concentrated surfactant ligands, the particle growth is carried out in a controlled fashion, and the resulting IONPs typically possess better crystallinity and higher magnetism than those made from the traditional methods. More importantly, such pyrolysis methods permit accurate control of the products, down to one nanometer size. This is significant because particle size and magnetic properties are tightly associated, and by tuning the reaction parameters, it is therefore possible to achieve a series of IONPs with controllable T2 contrast effects.
One disadvantage of pyrolysis-produced IONPs, however, is that the as-synthesized products are imbedded within a thick alkyl coating and are not water soluble. Many efforts have since been devoted to developing surface modification technologies that can confer conjugatability and water solubility to the particles. Despite their diversity, these surface modification techniques can be generally divided into two categories: ligand exchange and ligand addition. The prior refers to the strategy of using high affinity, hydrophilic ligands to replace the original hydrophobic coating. The latter refers to the use of amphiphilic materials, which are added to the particle surface by forming a bilayer structure with the existing alkyl coating.
IONPs with appropriate coatings can be easily coupled with drug molecules. For instance, the Zhang group coupled methotrexate (MTX), an anti-cancer drug, onto an aminated IONP surface. In vitro studies demonstrated that the particles, after internalizing into cells, accumulated in lysosomes, where the drug molecules were released due to the low pH and the presence of proteases. Hwu et al. reported on coupling paclitaxel (PTX) to IONP surfaces through a phosphodiester moiety at the (C-2′)-OH position. The average number of PTX molecules per nanoparticles was evaluated to be 83, and the release of the PTX was found to be more effective when exposed to phosphodiesterase. The Cheon group used meso-2,3-dimercaptosuccinic acid (DMSA) to modify IONPs, and used SMCC as the crosslinker to couple Herceptin antibody molecules onto the particle surface, although Herceptin was harnessed as a targeting agent rather than a therapeutic agent in the study.
Aside from covalent coupling, drug molecules can also be co-capsulated with IONPs into polymeric matrices. Jain et al. loaded doxorubicin (DOX) and PTX, along with oleic acid coated IONPs, into pluronic-stabilized nanoparticles. Similarly, Yu et al. loaded DOX into anti-biofouling polymer coated IONPs. When applied in a Lewis lung carcinoma xenograft model, such DOX loaded nanoconjugates showed better pharmacokinetics and therapeutic effects than DOX alone, presumably due to the anti-biofouling feature of the particles. Similarly, protein molecules have also been investigated as drug carriers. For instance, we developed a two-step coating strategy to yield human serum albumin (HSA) coated IONPs. With the excellent binding capacity of HSA, we expect that a range of lipophilic pharmaceuticals can be loaded into such nanoplatforms to yield theranostic agents.
It is known that small molecules can be loaded into porous nanostructures via physical absorption and, along that line, there have been efforts to achieve hollow iron oxide nanostructures. One such effort was reported by the Hyeon group. Starting from spindle-shaped β-FeOOH NPs made from FeCl3 hydrolysis, they performed a three step, so-called “wrap-bake-peel” treatment to achieve hollow IONPs. In a proof-of-concept study, they found that DOX could be loaded into such hollow nanoparticles via simple physical absorption and then released from the nanostructures in a sustained manner under physiologic conditions. A more recent report was given by the Sun group. By controlled oxidation and acid etching of Fe particles, they were able to achieve porous IONPs with a sizable cavity. They then loaded cisplatin into the cavities of the particles, and coupled Herceptin onto the particle surfaces to confer targeting specificity. The resulting conjugates showed selective affinity to ErbB2/Neu-positive breast cancer cells and a sustained cytoxicity attributable to the controlled release of cisplatin from the particle carriers.
Gene therapy has emerged as a crucial therapeutic avenue, where DNAs/RNAs are utilized as therapeutics to antagonize abnormal gene regulation. Unlike small-molecule-based therapeutics, DNAs/RNAs are negatively charged and, by themselves, have difficulty passing through the negatively charged cell membrane. In addition, the nucleases that are ubiquitous in living subjects may recognize and degrade DNAs/RNAs before they reach their targets and fulfill their task. It is in this scenario that nanoparticle-based delivery plays a significant role. In an ideal situation, the nanoparticle carriers can load the therapeutic genes, escort them to the diseased areas, facilitate their shuttling across the target cell membranes and, finally, release them intracellularly and fulfill their functions.
A milestone work along this line was reported by Medarova and his colleagues (Fig. 1). They coupled thiolated siRNA onto aminated dextran particles using N-succinimidyl-3-(2-pyridyldithio) propionate (SPDP) as a bridge compound. In addition, the near infrared dye Cy5.5 and myristoylated polyarginine peptide (MPAP), a membrane translocation peptide, were coupled to the particle surface. The probes were first equipped with siRNA that targets green fluorescence protein (GFP), and the conjugates were tested in a mouse model bearing bilateral 9L-GFP and 9L-RFP tumors. MRI and near-infrared fluorescence (NIRF) imaging were performed, and appreciable probe accumulation was found in both tumors. In contrast, optical signal drop was only observed in the 9L-GFP tumor, and not in the 9L-RFP tumor. siRNA-GFP was then switched to a therapeutic siRNA sequence that targets the antiapoptotic gene Birc5 (which encodes the protein survivin), and studied the therapeutic potential of the nanoconjugate in an LS174T human colorectal carcinoma xenograft model. RT-PCR results showed that such therapeutic particles induced an amazing drop in survivin transcript level of 97 ± 2%, accompanied by increased levels of tumor-associated apoptosis and necrosis.
In the above study, the particle accumulation in tumor was mediated by the EPR effect. It is also possible to add biovectors to nanoplatforms in order to impart site-specific delivery. For instance, Cheon group coupled siRNA and a PEGylated cyclic Arg-Gly-Asp (RGD) peptide onto magnetic nanoparticles. In a proof-of-concept study, MDA-MB-435 and A549 cells, which have high and negative integrin ανβ3 expression, respectively, were stably transfected with GFP and were incubated with nanoparticles loaded with GFP siRNA. The two cell lines showed distinct particle internalization rates, confirming that the particle uptake was mainly mediated by RGD-integrin interaction. Such differences in uptake resulted in a dramatic discrepancy in gene regulation efficacy. While the particles alone, with or with RGD coupling, showed no effect on knocking down GFP expression in A549 cells, a significant and concentration-dependent decrease in GFP expression was observed with MDA-MB-435 cells.
The magnetic properties of IONPs allow them to accumulate upon the summons of an external magnetic field. This feature has been utilized as a targeting mechanism to improve drug delivery efficiency, and related studies in animals and in human have been reported. In one such study, seven patients with metastatic breast cancer were infused with epirubicin-loaded IONPs (100 nm in diameter, at 0.5% of the estimated blood volume), and after that a magnetic field was established around the tumor. In about half of the patients, the magnetic field proved successful in directing the ferrofluid to the tumor to induce tumor regression. In a very recent effort, Namiki et al. screened cationic lipid coated IONPs, and found one formula, designated as LipoMag, that outperformed commercial PolyMag in both transfection and gene knockdown in all 13 tested cell lines (Fig. 2). The authors also screened and found the one sequence, siRNAEGFR#4, with the maximum percentage knockdown of EGFR mRNA. They then performed 2′-OMe modification on the uridine residues of the sense strand and yielded a modified sequence with similar knockdown effect but reduced cytokine induction compared with the parent sequence. Subsequently, they loaded this modified siRNA onto nanoparticles and evaluated their therapeutic potency in two gastric cancer models. A 50% reduction in tumor volume was observed in the therapeutic group after a 28-day treatment, along with other desired events, such as inhibition of angiogenesis and induction of apoptosis. Notably, such gene knockdown was only significant during application of magnetic fields at the tumor sites, as the lipid particles alone were not an effective delivery vehicle.
IONP can itself play an imaging/therapy dual role, due to its potential in hyperthermia. The underlying mechanism is that IONPs can act as antennae in an external alternating magnetic field (AMF) to convert electromagnetic energy into heat. This feature holds promise in tumor therapy for tumor cells that are more susceptible to elevated temperature than normal cells. In one example, phospholipid coated IONPs were injected into a subcutaneous tumor model in F344 rats, and were exposed to an AMF. The AMF in conjugation with IONPs raised the temperature of tumor above 43 °C and caused tumor regression, but had no effect on the control group where no IONPs were given. Also, Fab fragment of anti-human MN antigen-specific antibody was chemically anchored onto IONP surfaces and the INOPs were administrated systemically into tumor-bearing mice. The particles showed high tumor uptake, presumably due to an antibody-antigen interaction, and induced efficient tumor hyperthermia when exposed to an AMF. In another study, Zn-Pc, a photodynamic therapeutic (PDT) agent, was loaded onto CoFe2O4 nanoparticles, which showed better hyperthermia effects than IONPs. When evaluated in vitro with J774-A1 macrophage cells, a combined toxicity from both PDT and magnetohyperthermia was observed; however, more work needs to be done to elucidate the synergy of such combinational in vivo therapy.
Quantum dots (QDs) are light-emitting nanocrystals made from semiconductor materials. QDs are becoming an important class of biomaterials, because they possess unique optical properties that are unavailable from organic dyes or fluorescent proteins, such as being brighter, more photo- and chemical stable and possessing a narrow emission spectrum.
A unique feature of QDs is that their optical properties can be accurately adjusted by tuning their size and composition. The first generation of QDs was made of CdSe, CdTe and PbS, and by tuning of their sizes, gave rise to a series of nanomaterials that span most of the visible spectrum. However, such materials were found inefficient in in vivo applications due to the limited tissue penetration distances of visible light. To address this issue, exploits for QD formulas with near-infrared emission have been made, and those composed of CdTe/CdSe, Cd3P2, InAs/ZnSe and InAs/InP/ZnSe have been reported. An inorganic coating, such as ZnS, is usually added to the particle surface, and this has proved useful in enhancing the photoluminscent quantum efficiencies of the particles.
The synthesis of QDs is reminiscent of IONP preparation. In brief, appropriate organometallic precursors are heated in high boiling point organic solvent to initiate particle formation. Surfactants, such as trioctylphosphine (TOP) and trioctylphosphine oxide (TOPO), are utilized to control the particle growth. Similar to IONP synthesis, as-synthesized QDs are alkylated and are not water soluble. To confer water solubility, the most straightforward strategy is to add thiolated species to form disulfide linkage with the QD core or the ZnS shell. A host of small thiolated molecules have been investigated for such purposes, including mercaptoacetic acid, mercaptopropionic acid, mercaptosuccinic acid, dithiothreitol, glutathione, dithiothreitol (DTT) and cysteine. Interestingly, such old-school techniques have recently reentered scientific evaluations to prepare QDs with a minimized hydrodynamic size. A milestone finding has been that, when modified with cysteine, QDs can be rendered water soluble with a hydrodynamic size less than 5.5 nm and, when administrated systemically, these QDs have been found to be rapidly excreted via renal clearance, rather than being trapped in reticuloendothelial system (RES) organs, such as liver and spleen, as has been observed with most other nanoformulas. More recently, InAs/InP/ZnSe QDs have been coated with mercaptopropionic acid with an overall size of 8 nm, and intravenous injection in mice showed appreciable renal clearance that was confirmed by both NIRF imaging and urine sample assessment. One common concern of such a modification technique is the fragility of the disulfide bond. To strengthen the linkage, polydentate ligands, such as DHLA, oligomeric phosphines, cysteine rich peptides or multidentate polymers have been investigated, and they have been able to improve the longevity of the nanostructures. On the other hand, although it is S of the ZnS layer that is mostly utilized as a species-mounting site, it was reported that proteins/peptides with poly-histidine tags can be efficiently anchored onto QDs via strong Zn(II)-His interaction.
In addition to ligand-exchange, efforts have been made to investigate ligand-addition-based surface modification techniques, which involved a group of amphiphilic compounds, such as phospholipids, amphiphilic saccharides, acrylic acid polymers and others. One pioneering study was reported by Nie and his colleagues, who utilized a triblock copolymer (consisting of a polybutylacrylate segment, a polyethylacrylate segment, a polymethacrylic acid segment and a hydrophobic hydrocarbon side chain) to modify QDs and conjugated onto them a prostate-specific membrane antigen (PSMA) targeting antibody. When administrated into prostate cancer bearing mice, the nanoconjugates accumulated in the tumor area. This accumulation was attributed to both the EPR effect and specific antibody-antigen interaction.
QD based drug delivery is relatively less investigated, largely due to the innate toxicity of QDs. This problem is more prominent with the first generation QDs, where toxic Cd and Pb are frequently used in QD preparation. Recent advances in QD synthesis have led to the emergence of Cd-free QDs, such as InAs/ZnSe and InAs/InP/ZnSe. They are potentially more qualified candidate carriers, but relevant applications are so far limited. Nonetheless, Nurunnabi et al. reported on making QDs-Herceptin conjugates. The CdTe/CdSe QDs were made water soluble through the addition of PEG-10,12-pentacosadiynoic acid (PEG-PCDA), and the nanostructure was further stabilized upon UV irradiation, which crosslinked the coating shell and helped reduce the risk of Cd2+ leaking out from the core materials. The nanoparticles showed an efficient tumor targeting rate and impressive therapeutic effects when tested on an MDA-MB-231 tumor model. Park et al. co-encapsulated hydrophobic QDs and IONPs, along with DOX, into micelles formed with PEGylated phospholipid. Such conjugates were further coupled with a tumor-homing peptide F3, and were injected into an MDA-MB-435 xenograft model. Successful tumor targeting was observed by both optical and MR imaging modalities and was attributed to the mediation of the tumor-homing peptide. However, as a proof-of-concept study, the investigation stopped at the imaging level and no therapeutic studies were pursued.
In another interesting study, Bagalkot et al. produced a QD-aptamer(Apt)-DOX conjugate [QD-Apt(Dox)], and investigated its use for simultaneous cancer imaging, therapy and therapy monitoring (Fig 3). A10 RNA aptamer was coupled to the QD surface and was utilized as a biovector to target PSMA. The DOX loading, on the other hand, was achieved by its intercalation within the aptamer sequence. Interestingly, the fluorescence activities from QD and DOX were attenuated by their interaction with DOX and RNA, respectively, and were both in a quenched state in the nanosystem. However, when the particles were delivered into targeted tumor cells, DOX would be gradually released from the system, which, aside from initiating therapeutic functions, also led to the recovery of QD fluorescence. Similarly, Yuan et al. loaded MTX onto QD surfaces to induce photoluminescence quenching. The loading was achieved via simple reversible physical adsorption, which can be reversed when exposed to species with higher affinity, such as DNA. This coating change led to a restoration of the photoluminescence, which can be potentially useful to monitoring the delivery of drug molecules.
QDs may also function as gene delivery vehicles when modified with liopofectamine or other positively charged polymers. For instance, QDs have been encapsulated in poly(maleic-anhydride-alt-1-decene), and further surface-modified with dimethylamino propylamine to be positively charged. The resulting QDs outperformed polyethylenimine (PEI) by showing more efficient delivery and significantly reduced toxicity. Also with poly(maleic-anhydride-alt-1-decene) coated QDs, Gao et al. converted some carboxyls to tertiary amines with N,N-dimethylethylenediamine. The resulting particles, possessing two functional groups on the particle surface, afforded both steric and electrostatic interactions that were highly responsive to acidic endosome/lysosome organelles. When assessed as siRNA delivery vehicles, such QDs showed a 10–20 fold increase in silencing effect and a 5–6 fold decrease in toxicity when compared with other common delivery agents, such as lipofectamiane, JetPEI and TransIT. Covalently coupling siRNA to QDs has been investigated in addition to loading siRNA via electrostatic force. Derfus et al. covalently conjugated siRNA, along with a tumor-homing peptide F3, onto QDs. To study the feasibility of siRNA release, the coupling was achieved via two routes: with disulfide cross-linker and by forming a non-reducible, thioether linkage. Using eGFP as the target, it was shown that the former formula had greater gene silencing efficiency, which the authors attributed to the susceptibility of disulfide linkage in endosome/lysosome. More recently, QDs modified with amine functionalized polymer polydiallyldimethylammonium chloride (PDDAC) have been complexed with MMP-9-siRNA and utilized to modulate the activity of MMP-9, a main component of the blood brain barrier (BBB), in brain microvascular endothelial cells (BMVEC). The treatment caused an increase in the expression level of tissue inhibitor of metalloproteinase-1 (TIMP-1), a natural inhibitor of MMP-9 that functions to maintain the basement membrane integrity. In parallel, increases in collagen I, IV, V expression and a decrease in endothelial permeability were also observed.
QDs also have great potential in photodynamic therapy, where they act as either photosensitizers themselves or as carriers. The underlying mechanism for the first paradigm is that QDs can be activated by light and transfer the triplet state energy to nearby oxygen molecules to cause cell damage. Tsay et al. modified QDs with streptavidin and conjugated onto them biotinylated pDNA. They found the generation of reactive oxygen intermediates (ROI), through nitroblue tetrazolium (NBT) assay, when the QDs were photoactivated. Further investigation found that such ROI can elicit damage to purine and pyrimidine bases, as was confirmed by assays with base excision repair enzymes, such as formamidopyrimidine glycosylase (Fpg) and endonuclease III (Endo III). Compared with small-molecule-based photosensitizers, such QD based PDT offers advantages, such as better chemical stability, water solubility, and (for NIRF QDs) less optical interference with biological tissues. The drawback, however, is that the quantum yield of such QD-generated singlet oxygen is typically less than 5%, much lower than that of classic photosensitizers (40–60%).
QDs have also been investigated as carriers for PDT agents. Hsieh et al. conjugated QDs with Ir-complex. Samia et al. conjugated QDs with phthalocyanine (Pc4). Shi et al. used meso-tetra(4-sulfonatophenyl)porphine dihydrochloride (TSPP) as a photosensitizer and coupled it onto QD surfaces. Note that, in the latter two cases, the QDs worked as drug carriers and were not directly involved in the photodynamic therapy. In Samia’s system, however, the QD also worked as an energy hub, which transferred energy to phthalocyanine (Pc4) to activate its PDT function.
The use of QDs for biomedical application arises from their unique physicochemical properties, including broad absorption spectra, size-dependent narrow and stable emissions spectra, photostability, and the ability to serve as scaffold for additional agents such as targeting ligands and therapeutic drugs. In addition, there are numerous unexplored possibilities to expand the use of QD-based theranostic agents based on its multiplexing and FRET capabilities. However, there is an urgent need to understand the metabolism of QDs in the body and address the heavy metal related toxicity issues.
Gold nanoparticles (Au NPs) possess many unique features and have been investigated in a variety of imaging related arenas, such as in computed tomography (CT), photoacoustics and surface-enhanced Raman spectroscopy (SERS). Synthesis of Au NPs has been well established, and those in the forms of spheres, cubes, rods, cages and wires can now be acquired with accurate quality control and in large quantity. Such morphology control is important as it greatly influences the physical properties of the products and in turn affects their role as imaging probes. For instance, 10 nm spherical Au NPs have characteristic surface plasmonic absorption at around 520 nm. Increasing the particle size leads to some, but not dramatic, red-shift of the particle absorption spectrum--the maximum absorption of 48.3 and 99.4 nm Au NPs are at 533 nm and 575 nm, respectively. Changing the nanoparticle shape to rod-like, on the other hand, can push the absorption to the NIR region (650–900 nm), which suggests their role as probes in photoacoustic imaging or mediators in photothermal therapeutics.
Due to the strong interaction between thiol and Au, the surface modification of Au nanostructures is overwhelmingly conducted via addition of thiolated species. Typically, a bifunctional compound is utilized, which has its thiol termini immobilized onto the particle surface, while leaving carboxyl/amine termini exposed for conjugation with functional entities. Alternatively, biomolecules can be pre-thiolated and loaded as a whole onto particles. For instance, thiolated DNA oligos have long been utilized for stabilizing Au NP colloids and the resulting conjugates have been investigated as gene therapy agents. Unlike surface modification of QDs, where the fragility of disulfide binding is a concern, a monodentate thiol is generally regarded sufficient to induce stable ligand anchoring of Au. As a matter of fact, monodentate ligands can be more advantageous than multidentate species to allow higher loading capacity. This is evidenced by a study where two oligonucleotides -- one tetrathiol modified and the other monothiol modified--were loaded onto separate 13 nm Au NPs. Although the tetrathiol oligos and the monothiol oligos shared the same sequence, only 45~50 tetrathiol strands can be loaded onto each particle, compared to 110~120 for monothiol oligos.
In the nanorod synthesis, cetrimonium bromide (CTAB) is widely used. It works by forming a thick double-layer around the side wall of the nanorods, and causes a longitudinal growth of the nanostructure. Thus, most thiol-Au interaction occurs at the ends of the nanorods, since the side wall is hard to access. In order to achieve a better loading rate, a layer-by-layer deposition approach is sometimes utilized. This method takes advantage of the highly positively charged CTAB coating and loads alternatively charged species to the particle surface. In such a manner, large functional molecules, such as antibodies, can be directly immobilized onto the outer layer of the particle surface with electrostatic interaction.
As already mentioned, Au-thiol chemistry is favored to load functional entity onto Au NPs, and a number of therapeutics have been loaded in such manner. For instance, PTX was modified at its C-7 position and covalently coupled to 4-mercaptophenol modified Au nanoparticles. The resulting conjugates demonstrated better therapeutic effects than MTX alone both in vitro and in vivo, which was likely due to the “concentrated effect” and an improved pharmacokinetics of the conjugates. Similarly, protein based pharmaceutics have been loaded onto Au NPs. Tumor necrosis factor (TNF), for instance, was coupled to PEGylated Au nanoparticles, and the resulting conjugates showed better therapeutic efficacy and less toxicity than native TNF.
In addition to thiol-Au association, other ways of anchoring pharmaceutics have also been investigated. Pokharkar et al. used chitosan as a reducing agent and coating material to make Au nanoparticles. The resulted chitosan-Au nanoparticles were highly positively charged and were found highly efficient in loading insulin via electrostatic interaction (53%). Such conjugates were studied in a diabetic model to control postprandial hyperglycemia. Two hours after administrating these insulin loaded Au NPs to diabetic rats, a drop in their blood glucose level was observed, of 30.41% and 20.27% for oral (50 IU/kg) and nasal (10 IU/kg) administration, respectively. Cheng found that a PDT agent, Pc4, can be directly adsorbed onto PEGylated Au nanoparticles with high efficiency. The Au nanoparticles worked well as a drug carrier, reducing the time for sufficient Pc4 delivery to less than 2 h, compared to that of 2 days for free drug. Similar work with zinc-Pc was reported by Hone et al.. Prabaharan et al. utilized an amphiphilic-block-copolymer-coated Au NP formula for tumor targeting and drug delivery. Such a nanostructure consisted of a Au NP core, a hydrophobic PASP inner shell, and a hydrophilic, folate-conjugated PEG outer shell (PEG-OH/FA). DOX was covalently conjugated onto the hydrophobic inner shell by acid-cleavable hydrazone linkage, with a loading level of 17 wt%. Such a nanosystem is interesting for possessing both a tumor targeting mechanism (folate on the outer layer) and an intracellular drug release mechanism (hydrazone linkage of DOX on the inner layer).
Au NPs have also been converted to polyelectrolytes and studied as gene delivery carriers. Previously, Au NPs have been functionalized with alkylated quaternary ammonium to load plasmid DNA. In vitro, such nanocarrier showed benefits such as protecting DNA from enzymatic digestion and GSH activated release. Later, Klibanov et al. used branched PEI to confer gene loading capacity to Au NPs. Under optimized conditions, transfection potency can be increased by 12 times, compared to the parent polymer. In addition to electrostatic forces, therapeutic genes can also be loaded onto Au nanoparticles through covalent linking. As mentioned above, thiolated antisense DNA oligos can be directly loaded onto Au nanoparticles with high efficiency. In a cellular study with eGFP-expressing C166 cells, Au NPs loaded with antisense DNA showed a high translocation rate and a prominent gene knockdown efficiency.
The unique surface plasmon resonance feature of Au NPs makes them candidate materials in photothermal therapy. The idea is to use Au NPs to serve as energy transducers, which, when concentrated in tumor areas and upon laser irradiation, can convert light into heat and kill adjacent cancerous cells. Compared with conventional drug delivery, such a treatment paradigm is active only within the limited illumination area, therefore minimizing normal tissue damage. As mentioned above, spherical Au NPs, with a characteristic absorption at 500–600 nm, are not appropriate materials for such an application. Changing the configuration to a nanorod, nanocage or nanoshell, on the other hand, can shift the absorption to the NIR region, and reports on gold-nanostructure-based hyperthermia are accumulating. For instance, the Xia group demonstrated recently that PEG coated Au nanocages can accumulate in a U87MG xenograft model, and when exposed to NIR light, can increase the tumor surface temperature to 54 °C within 2 min. Li coupled α-melanocyte-stimulating hormone (MSH) analog, [Nle4, D-Phe7]α-MSH (NDP-MSH), onto Au nanoshells and administrated such conjugates to a B16/F10 melanoma model (Fig 4). Careful histological examination found high accumulation of particles in the tumor and confirmed that such homing was mediated by NDP-MSH. Efficient ablation of B16/F10 melanoma was found in the tumor exposed to laser illumination, but not in the contralateral tumor where no illumination was given. Successful photothermal therapy was validated histologically and, more interestingly, [18F] fluorodeoxyglucose (18F-FDG) PET found a remarkable decrease in tumor uptake, reflecting a metabolic activity drop upon photothermal therapy. More recently, the same group reported the use of Au nanoshells as light-controllable siRNA carriers (Fig 5). The particles were conferred with tumor targeting specificity by imparting folic acid to the particle surface. The siRNA, with a sequence that targets NF-κB P65, was pre-thiolated at the 5′ end of the sense chain and was introduced to the particles surface via thiol-Au interaction. It was concluded that the stable thiol-Au association could carry the siRNA payload on the particle surface, even after cell uptake, but would be destroyed when exposed to NIR light irradiation, which damaged the endolysosomal membrane and led to a release of siRNA into cytoplasms. This mechanism of action was confirmed by the observation of light-inducible siRNA release and subsequent NF-κB P65 downregulation both in vitro and in vivo. Furthermore, it was found that the downregulation of NF-κB P65 resulted in increased sensitivity to chemotherapy, as evidenced by an improved therapeutic index when such photothermal therapy was combined with irinotecan treatment.
With unique characteristics, such as strong surface plasmon absorption, stability, biosafety, and ease of modification, AuNP have long been exploited as a candidate material for building up functional agents for both imaging and therapy applications. One obvious disadvantage is its high cost of production, which may cast a shadow over applications that are otherwise of bright clinical perspectives. Also notably, while the thiol-Au chemistry is convenient and is generally regarded as sufficient to prepare stable AuNP conjugates, it can be an issue when exposed to a reducing environment, such as glutathione (GSH) that is of high concentration in a living subject. While more and more investigations are to be performed at in vivo level, new chemistry that are able to result in more stable conjugates can be highly favorable.
Carbon nanotubes (CNTs) have found potential applications in Raman and photoacoustic imaging and have been studied as drug carriers by a number of research groups. CNTs have a graphite-like structure, which is inert and inhibitive to most conjugation chemistry. To address this issue, researchers have tried to apply extreme oxidative conditions, which generate defects on the CNT surface that can be utilized as mounting sites. For example, after treating single walled nanotubes (SWNTs) in refluxing 2.5 M HNO3 for two 36 h periods with intervals of 30 min sonication, the SWNTs can be made water soluble due to the generation of multiple carboxyl groups on the nanotube surface. Subsequent conjugation can be applied to covalently link molecules onto those nanotubes by forming amide bonds. Alternatively, it has been reported that azomethine ylide and its derivatives can anchor onto CNT surface through 1, 3-dipolar cyclo-addition. A group of functional molecules, including organic dyes, peptides and antibiotics, have been conjugated to the CNT surface in such a manner.
The hydrophobic and aromatic nature of CNT surface also encourages non-covalent molecule anchoring. Certain amphiphilic compounds, such as sodium dodecyl sulfate (SDS), triton-X-100, CTAB, sodium dodecylbenzene sulfonate (SDBS) and sodium dodecane sulphonic acid (SDSA) have been used to disperse CNT in aqueous solution. One interesting finding is that SDBS is much more efficient than its analog SDS in producing CNT suspension, which has been attributed to the aromatic ring in SDBS that interacts with the CNT surface and further lowers the energy. Indeed, aromatic compounds, such as 1-pyrenebutanoic acid succinimidyl ester, PS-b-PAA and 1-pyrenepropylamine hydrochloride have proved effective in modifying CNTs. Along this line, there have been recent efforts to use DNA oligos to modify CNTs. With multiple bases forming π-π interactions with the graphite surface, the DNA oligos can wrap around the CNT in a helical manner, and disperse the conjugates with the sugar-phosphate backbones exposed to the solution.
Among the many ligands that are available to functionalize the CNT surface, PEGlyated phospholipids might be one of the most studied due to its effectiveness and biocompatibility. The Gambhir group conducted a four-month careful monitoring of mice intravenously administrated phospholipid coated CNTs. Even though some CNTs were found trapped in the liver and spleen without degradation, no evidence of toxicity was found from the survival, clinical, and laboratory parameters, necropsy and tissue histology studies. Although it is still debatable whether CNT is safe or not for in vivo applications, such studies at least suggested a possibility that inconsistent observations in CNT toxicity studies could be caused by the coating materials, rather than the CNTs themselves.
Carbon nanotubes can be taken up by cells, and this feature has inspired a wave of research efforts to investigate their potential role in drug delivery. The detailed mechanisms underlying such efficient cell penetration are unclear. As a matter of fact, CNTs can be internalized by cells via different routes, if surface coatings are different and with appropriate coatings, as both endocytosis and passive diffusion have been reported to be responsible for carbon nanotube uptake.
Previously, Prato et al. coupled MTX onto 1, 3-dipolar cycloaddition functionalized CNTs. Similar CNT formulas with multiple amine termini were used to load and deliver DNA plasmid. However, at least in the MTX case, it was found that the covalent amide bond was unfavorable to eliciting intracellular drug release, since the conjugates showed no enhancement in therapeutic efficacy as compared to MTX alone. The Dai group applied phospholipid-CNT conjugates in both imaging and therapy. For example, they coupled siRNA to CNTs via a disulfide bond, which was susceptible to enzymatic breakage in the endolysosome. This CNT transporter showed high transfection efficiency, outperforming lipofectamine in inducing RNAi. Later, they used the same nanostructure and successfully shuttled siRNA into human T cells and primary cells, which are generally regarded as hard to transfect with conventional cationic liposome-based transfection agents. The same group also reported the coupling of either Pt(IV) prodrug or PTX onto PEGylated CNTs to improve the pharmacokinetics and therapeutic effects. In the case of PTX, a branched-PEG was used for phospholipid PEGylation, which was found to be advantageous over single-chain PEG in bringing extra stability to CNTs (Fig 6). PTX was coupled through a cleavable ester bond to the nanotube surface and the construct was then tested in a murine 4T1 breast cancer model. The conjugates showed a 10-fold increase in tumor homing than PTX alone, which was attributed to the prolonged circulation half-life of the nanoformula. Consequently, this formula showed better tumor suppression outcome than clinically used Taxol.
Aromatic stacking can also serve as a drug loading mechanism. DOX, for instance, was loaded onto SWNT via this route at a remarkably high efficiency of 4g DOX/g nanotube. The interaction is pH dependent, suggesting a way of unloading the payloads in acidic endolysosome and tumor micro-environments. More recently, such DOX loaded nanotubes have been evaluated in SCID mice bearing Raji lymphoma xenografts, which showed greater therapeutic efficacy and less toxicity compared to an equimolar amount of free DOX (Fig 6).
The strong optical absorbance of CNTs in the NIR region has made it a promising tool in photothermal therapy. Previous studies have shown that, when irradiated by NIR light, CNTs that were internalized in cells were capable of triggering endosomal rupture and cell death. More recently, Moon et al. demonstrated in a human epidermoid mouth carcinoma model that the combined treatments of PEGylated SWNT and NIR irradiation led to the eradication of tumors with no observation of recurrence over 6 months. Also, Ghosh et al. used DNA to encapsulate CNT, which, according to the authors, can lead to improved heat emission efficacy. When injected in a PC3 xenograft model and irradiated, the conjugates induced complete tumor eradication. However, it is worth pointing out that in both cases the nanotubes were injected intratumorally rather than systemically.
Overall, CNTs have unique physical and surface features and are a key player in the field of nanomedicine. Aside from loading drug molecules, other nanopartices such as IONPs or AuNPs can be easily tethered on CNTs to further enrich the nanoplatform. However, their non-biodegradablity remains to be a great concern. It is debatable whether residual CNTs can cause chronic and longitudinal damage to the host. In addition, there is a lack of a standard protocol to prepare CNTs of high purity and at large scale, which is mandatory for a clinical translation.
Silica is generally regarded as a safe material, and has been previously used as surgical implant. It is well documented, as well, that accurate size and morphology control are achievable in the synthesis of silica nanoparticles. Generally, silica nanoparticles themselves do not have characteristics for imaging. Instead, they afford an excellent platform that allows facile loading of a broad range of imaging and therapeutic functions, making them a good candidate for theranostic purposes.
Silica nanoparticles can be formed by hydrolysis and condensation of tetraethyl orthosilicate (TEOS). To introduce functional groups onto the particles, it is very common to use aminopropyltrimethoxysilane (APS) or mercaptopropylmethoxysilane (MPS) as co-precursors, which co-coagulate with the TEOS matrix to bring amine or thiol groups to the particle surface. A functional molecule can be easily imported into the nanosystem during the particle formation, if it is pre-coupled with APS/MPS. Both organic dyes and Gd-DTPA complexes have been coupled and integrated into silica particle matrix through this approach to yield optically or magnetically active agents.
Aside from small molecules, nanoparticles can also be easily incorporated into silica matrices, and reports on using such technology to encapsulate IONPs, Au NPs and QDs have been well documented. Moreover, several functionalities can be encapsulated into a single silica particle simultaneously. For example, Nie et al. utilized silica to encapsulate both IONPs and QDs, creating a hybrid that retained both magnetic and optical properties. Koole et al. reported a dual-function core-shell-shell (CSS) nanoparticle, where silica nanoparticles were loaded with both QDs and Gd complexes.
Drug molecules can be easily introduced into silica nanoparticles during particle formation. Roy et al. incorporated 2-devinyl-2-(1-hexyloxyethyl)pyropheophorbide (HPPH), a hydrophobic photosensitizing anticancer drug, into silica matrices. It was demonstrated that the HPPH is more fluorescent in the silica matrices than in the free form, and can efficiently kill cancer cells when irradiated with a laser. Recently, the same group co-encapsulated HPPH and a two-photon absorbing dye, 9,10-bis[4′-(4″-aminostyryl)styryl]anthracene (BDSA), into silica nanoparticles. It was shown that BDSA can efficiently up-convert the NIR light and intrapartically transfer the energy to HPPH to activate the latter’s PDT function.
A breakthrough in silica nanoparticle preparation is that they can be made in mesoporous form with accurate pore size control. Such mesoporous nanostructures, consisting of hundreds of empty channels and a large surface area (>900 m2/g), are excellent reservoirs for small molecules and they hold great promise in drug delivery. Many methods have been reported, chemically and physically, to achieve such nanostructures. In a typical chemical preparation, n-alkyltrialkoxysilane or other surfactants are mixed with other precursors and are incorporated into the matrices during particle formation. These surfactants will be removed later from the nanostructure via post-synthesis solvent extraction or calcinations to yield the mesoporous structure. With the molecular-sieve structure, such mesoporous silica nanoparticles can load many small molecule pharmaceuticals via simple physical interaction. Furthermore, technologies have been developed to cap the mesopores after drug loading to inhibit premature drug release. For instance, mesoporous silica nanoparticles were loaded with PTX, and the mesopores were subsequently capped with Au NPs. Such Au NP capping was designed to be photolabile and can be uncapped to release guest molecules when photoirradiated. Similarly, activatable gatekeepers based on QDs, IONPs, courmarin and diethylenetriamine have also been reported.
More recently, Park et al. reported on the preparation and application of luminescent porous silicon nanoparticles (LPSiNPs) (Fig. 7). The nanoparticles were made in physical ways. A typical preparation consisted of HF etching of single-crystal silicon wafers, lifting-off of the porous silicon film, ultrasonication, filtration of the formed particles through a 0.22 μm filtration membrane and finally activation of luminescence in an aqueous solution. The luminescence was generated by quantum confinement effects and the defects localized at the Si-SiO2 interface. Such luminescent porous nanoparticles were loaded with DOX and their drug release and cytotoxicity were studied in vitro. One unique feature of such silica nanoparticles is that they can self-destruct in vivo and be renally cleared within a relatively short period of time, therefore reducing the risk of their getting trapped in and causing damage to normal organs.
In the current article, we have highlighted some nanoplatforms that are currently under intensive investigation for the build-up of theranostic agents. All of the nanoplatforms discussed here have gone through years of development and allow facile and reliable function docking. These nanoparticles can possess unique optical or magnetic properties and have been previously studied in the imaging setting and have achieved many successes. These have laid the foundation for the current exploits, since the imaging probes can be easily upgraded when loaded with appropriate therapeutics. It has been shown that therapeutics of various forms, including those that are small molecule, protein and nucleotide-based, can be conveniently tethered onto nanoplatforms. The great capacity even allows the loading of a second or third functionality, a feature which encourages the formation of an all-in-one nanosystem with comprehensive features.
However, despite the fast progress, there has so far been no nanoparticle theranotics that is so developed to meet clinical standards. Each nanoplatform has its own promises and advantages, but meanwhile, has its disadvantages to be overcome. These include the toxicity of QDs, the cost of gold nanoparticles, the intrinsic low sensitivity of IONP as MRI contrast probes, the nonbiodegradable nature of CNTs, and the oversize of silica nanoparticles, to name a few. Efforts to address each of these issues are going on, and should remain the focus in the future studies. Furthermore, as drug carriers, specific targeting is a topic that can never be over addressed. We have discussed the efforts of introducing a targeting motif onto nanoplatforms to achieve an improved targeting profile. While this route should continue to be exploited, it is also advisable to address the issue by using nanoparticle-based hyperthermia, where treatment only occurs in a confined region and allows minimal normal tissue damage.
In addition to achieving and validating the nanoscale integration of imaging and therapeutic functions, it is of significant importance to demonstrate the benefits and synergy of such a combined approach. In theory, a NP-based theranostic agent can deliver therapeutics to a diseased area and can use its imaging function to improve diagnosis and to monitor therapeutic response. Despite the promise, the related proofs are so far inadequate and should be the main focus of the next stages of investigation.
This work was supported by the Intramural Research Program, NIBIB, NIH. S.L. is partially funded by the NIH/NIST NRC fellowship. We acknowledge Dr. Henry S. Eden for proofreading this manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.