Skip main navigationJournal menuClose Drawer MenuOpen Drawer Menu
Skip main navigationJournal menuClose Drawer MenuOpen Drawer Menu
Journal of The Royal Society Interface
You have accessReview articles

Optical imaging-guided cancer therapy with fluorescent nanoparticles

Published:16 September 2009https://doi.org/10.1098/rsif.2009.0243

    Abstract

    The diagnosis and treatment of cancer have been greatly improved with the recent developments in nanotechnology. One of the promising nanoscale tools for cancer diagnosis is fluorescent nanoparticles (NPs), such as organic dye-doped NPs, quantum dots and upconversion NPs that enable highly sensitive optical imaging of cancer at cellular and animal level. Furthermore, the emerging development of novel multi-functional NPs, which can be conjugated with several functional molecules simultaneously including targeting moieties, therapeutic agents and imaging probes, provides new potentials for clinical therapies and diagnostics and undoubtedly will play a critical role in cancer therapy. In this article, we review the types and characteristics of fluorescent NPs, in vitro and in vivo imaging of cancer using fluorescent NPs and multi-functional NPs for imaging-guided cancer therapy.

    1. Introduction

    Cancer is one of the major causes of mortality and morbidity in the world and the incidence of cancer continues to increase. The common cancer treatment methods include surgery, chemotherapy and radiation therapy; however, current therapies are limited by ineffective early diagnosis, insufficient drug concentrations reaching the tumour, systemic toxicity, adverse effects of drugs and lack of capability to monitor therapeutic responses. Thus, the important issues in improving treatment regimens are (i) development of advanced imaging technologies for early diagnosis and to track therapeutic responses; (ii) utilization of targeting moieties to specifically and efficiently deliver drugs to tumour tissues; and (iii) discovery in cancer biology and genetics leading to new knowledge for cancer treatment.

    Nanotechnology is a new field of interdisciplinary research to fabricate materials with nanoscale dimensions between 1 and 1000 nm (Ferrari 2005). Materials at the nanometre scale have novel optical, electronic, magnetic and structural properties compared with the same materials at bulk volume, making them attractive for use in cancer diagnosis and therapy. Recent research has developed a number of nanoparticles (NPs), such as metal, semiconductor and polymeric particles, to be used as imaging probes, delivery vehicles, and even as multi-functional NPs (Liu et al. 2007a; Wang et al. 2008; Riehemann et al. 2009). NP-based drug-delivery systems based on chitosan, polyethyleneimine (PEI), liposomes, micelles and silica NPs offer the potential to optimize drug delivery while reducing drug side effects (Sinha et al. 2006; Cho et al. 2008). There are also several types of NPs used in optical molecular imaging in cancer diagnosis, such as organic dye-doped polymer and liposomes, quantum dots (QDs) and upconversion NPs (UCNs; Licha & Olbrich 2005; Santra et al., 2005; Grodzinski et al. 2006; Rao et al. 2007). More importantly, NPs are capable of combining different modalities (targeting, imaging, drug delivery and sensing) on one particle, which leads to multi-functional NPs for simultaneous tumour imaging and treatment (Torchilin 2006; Sanvicens & Marco 2008; Suh et al. 2009). With the engineered multi-functional NPs, imaging-guided cancer therapy can be realized.

    In this article, we review the types and characteristics of fluorescent NPs, in vitro and in vivo imaging of cancer using fluorescent NPs and multi-functional NPs for simultaneous tumour imaging and treatment.

    2. Fluorescent nanoparticles

    Optical imaging is the latest trend in imaging-guided therapy that involves the detection of light photons transmitted through tissues. It can non-invasively monitor the progression of disease and therapy. Conventional fluorophores such as fluorescent dyes, bioluminescent proteins and fluorescent proteins were used initially. But the recent advancements in the development of fluorescent NPs have made them potential candidates for imaging-guided therapy and they have a lot of advantages over their predecessors.

    2.1. Organic dye-doped nanoparticles

    Recently, there has been a surge in the development of NPs doped with organic dyes for imaging. Nanoencapsulation of the organic dyes makes them more stable and amplifies the signal considerably. The NPs are usually made of silica, but sometimes other polymers like poly(d,l-lactic-co-glycolic acid) (PLGA; Suzuki et al. 2008) are also used. These NPs have been doped with many organic dyes like IRG-023 Cy5, fluorescein isothiocyanate (FITC), rhodamine B isothiocyanate (RITC), etc. (Liu et al. 2006; He et al. 2007; Shi et al. 2007b; Suzuki et al. 2008). Labelling of NPs with a combination of dyes has also been reported (Wang et al. 2005a; Gao et al. 2007).

    The organic dye-doped NPs are usually synthesized by two main methods namely the Stober method and the microemulsion method. The size varies from 2 to 200 nm and can be controlled. The NPs produce light of high intensity due to the large number of dye molecules within each particle and they are quite photostable. The photostability is mainly due to the polymer coating that prevents the penetration of oxygen, thereby reducing the bleaching (Zhou & Zhou 2004). Many of these NPs exhibit good biocompatibility and water solubility and universal bioconjugation strategies can be used for attaching biomolecules to them. The versatile silica chemistry is used for bioconjugation through functional groups such as thiol, amino and carboxyl groups. Interactions between avidin and biotin are also employed (Tapec et al. 2002).

    2.2. Quantum dots

    QDs are semiconductor crystals with sizes in the nanometre range. They are composed of elements from groups II to VI, III to IV or IV to VI from the periodic table. The size of the QD is usually from 2 to 10 nm, which gives them special properties not seen at a macro-level owing to the effect of quantum confinement. QDs have a broad absorption spectrum, i.e. they can be excited by a wide range of wavelengths and they have a narrow emission spectrum. They also have extensive tunability whereby the emission wavelength can be controlled by the size of the QDs. Multi-coloured QDs that can be excited by a single wavelength are very useful in cellular imaging with multiple labels.

    QDs are usually synthesized by heating the precursors dissolved in organic solvents at high temperatures of about 300°C (Dabbousi et al. 1997; Talapin et al. 2001; Reiss et al. 2002; Bang et al. 2008). The size of the QDs can be varied by varying the concentration of the precursors and the crystal growth time. The nanocrystals thus formed have a hydrophobic core and are thus insoluble in water. So various surface modifications such as silica encapsulation, ligand exchange, conjugation to mercaptohydrocarbonic acids, dithiothretol and oligomeric ligands are carried out to make them soluble in water, which is essential for biological applications (Gerion et al. 2001; Pathak et al. 2001; Kim & Bawendi 2003; Nann & Mulvaney 2004). Conjugation of biomolecules on the surface of the QDs is usually by physical adsorption and electrostatic interactions (Lakowicz et al. 2000; Mattoussi et al. 2000; Jaiswal et al. 2003) but covalent coupling with linkers (Goldman et al. 2001; Winter et al. 2001; Parak et al. 2002) is also used routinely.

    QDs are very bright, photostable and thermostable. They are quite resistant to photobleaching and can be used for in vivo tracking for extended periods of time. Toxicity of these QDs has always been in question and cytotoxic studies of CdTe and CdSe QDs were reported (Shiohara et al. 2004; Lovric et al. 2005). However, the cytotoxicity is dependent on the dose, mode of particle preparation, surface coating, etc. The coating of the particles with hydrophilic polymers and ZnS prevents the leaching of the toxic elements such as cadmium and selenium, thereby reducing the toxicity considerably. Thus, QDs are more stable for long times in in vivo imaging than the dye-doped NPs and have been used widely (Dubertret et al. 2002; Gao et al. 2004; Jaiswal et al. 2004).

    To overcome the problems of using UV light as the excitation source for the visible QDs, near-infrared (NIR) QDs were developed, which can be excited by NIR light (Iyer et al. 2006; Yong 2009). The use of UV light as a source causes damage to biological components, generates singlet oxygen, has low penetration depths, high background autofluorescence, etc. Thus, NIR QDs overcome all these problems and are more efficient and suitable for in vivo and real time fluorescence imaging.

    2.3. Upconversion nanoparticles

    UCNs are phosphors that absorb light in the NIR region and emit in the visible region. The UCNs are usually synthesized with host lattices such as LaF3, YF3, Y2O3, LaPO4, NaYF4 codoped with trivalent rare earth ions such as Yb3+, Er3+ and Tm3+ (Wang et al. 2006; Guo & Qiao 2009; Lisiecki et al. 2009). The host lattice with crystals of yttrium fluoride, such as NaYF4 particles codoped with Er3+/Yb3+/Tm3+, was found to be the most efficient in the process of upconversion (Suyver et al. 2005, 2006; Schafer et al. 2007). The rare earth lanthanides doped in crystal centres of UCNs act as absorber ions and emitter ions. The absorber ion (e.g. Yb3+) is excited by an infrared light source and then transfers this energy non-radiatively to the emitter ion (e.g. Er3+ or Tm3+) that radiates a photon. This class of NPs has gained popularity in recent years and has been used for various biological applications such as imaging (Chatterjee et al. 2008; Jiang & Zhang 2008), photodynamic therapy (Chatterjee & Yong 2008; Ungun et al. 2009), FRET studies (Wang et al. 2005b; Kuningas et al. 2006), biomolecules delivery (Jiang et al. 2009), immunoassays (Kuningas et al. 2007; Ukonaho et al. 2007), immunocytochemistry (Zijlmans et al. 1999), immunochromatography (Hampl et al. 2001), microarray applications (van de Rijke et al. 2001) and DNA detection (Corstjens et al. 2001; Wang & Li 2006).

    UCNs are usually synthesized at very high temperatures or in organic solvents in the presence of surfactants. These processes produce NPs that are insoluble in water, non-biocompatible and lack functional groups for conjugation to biomolecules. Wang et al. (2006) developed a more efficient hydrothermal synthesis process, where the NPs were synthesized with a PEI coating that makes the particles biocompatible. Other strategies using surfactants such as poly(vinyl pyrrolidone) and oleic acid and encapsulating with polystyrene were also used to obtain enhanced stability, solubility and functionality (Li & Zhang 2006; Li et al. 2008; Qian et al. 2008a).

    Fluorescent UCNs have excellent photostability, chemical stability, low toxicity and are biocompatible (Chatterjee et al. 2008). The main advantages of these NPs over visible QDs and organic dye doped NPs is their ability to be excited in the NIR region, where autofluorescence is minimal, tissue penetration is maximum and there is minimum photodamage. They also do not exhibit photoblinking, which is a phenomenon observed in QDs. NIR QDs have many comparable properties to UCNs like high penetration depth but they are comparatively costlier than UCNs and more toxic. The upconversion fluorescence output of UCNs is also higher than that of QDs by seven orders of magnitude (Heer et al. 2004). Biodistribution of UCNs has been studied and they are cleared from the body mostly by 7 days (Jalil & Zhang 2008). Thus, fluorescent UCNs are becoming a class unto themselves with a lot of advantages over their predecessors and are excellent for biological applications. A comparison of organic dye-doped NPs, QDs and UCNs is presented in table 1.

    Table 1.Comparison of organic dye-doped NPs, QDs and UCNs.

    Collapse
    parameterorganic dye-doped NPsQDsUCNs
    size50–500 nm2–10 nm50–200 nm
    autofluorescencehighhighlow
    light penetration depthmedium/highmedium/highhigh
    photodamagemedium/lowmedium/lowlow
    cytotoxicitymediumhigh/mediumlow
    biocompatibilitygoodgoodgood
    photostabilitylowhighhigh
    excitation wavelengthUV/vis/NIRUV/NIRNIR
    costlow/mediumhighlow
    excitation radiation toxicitymedium/lowmedium/lowlow
    multiplexing assaysn.a.goodgood

    3. In vitro imaging of cancer

    To realize the benefits of early cancer diagnosis, highly sensitive and specific assays for biomarkers are needed: a biomarker is an indicator of a particular state, either normal or diseased state of an organism. By critically defining the implication among these biomarkers, it is possible to diagnose and prognosticate cancer based on a patient's molecular profile, leading to personalized and predictive medicine. Within the last several years, lots of articles have described the ability of fluorescent NP probes to accurately and quickly quantify biomarkers on cancer cells or tissue specimens, allowing a non-invasive detection for cancer.

    Some receptors or antigens on cell plasma membrane have been studied as cancer biomarkers. Selection of the appropriate receptors or antigen on cancer cells is important for specific cancer diagnosis or receptor-mediated delivery of therapeutic agents. The ideal targeted antigens should have abundant and unique expression on cancer cells, but undetectable or low expression on normal cells. The targeted NPs can be internalized after binding to the antigens, increasing the intracellular concentration of drugs (Goren et al. 1996). Plasma membrane antigens on living cells such as integrins (Lieleg et al. 2007; Garg et al. 2009), folate receptors (Zhang & Huang 2006; Kim et al. 2008a; Jiang et al. 2009), transferrin receptors (Qian et al. 2007, 2008b) and erbB2/HER2 (Wu et al. 2003; Lidke et al. 2004; Wartlick et al. 2004; Cirstoiu-Hapca et al. 2006; Tan et al. 2007; Anhorn et al. 2008) have been specifically recognized and tracked with fluorescent NP bioconjugates. On the other hand, the choice of targeting moieties that are modified on the NPs is also important to successfully bind the cancer cell and trigger receptor-mediated endocytosis. A variety of affinity agents, such as monoclonal antibodies (mAb; Santra et al. 2001; Cirstoiu-Hapca et al. 2006; Tan et al. 2007), receptor ligands (Zhang & Huang 2006; Huang et al. 2007), recognition peptides (Lagerholm et al. 2004; Ruan et al. 2007) or aptamers (Bagalkot et al. 2007), have been used to facilitate the uptake of carriers into target cells. A review outlined the major cancer targets for NP systems (Byrne et al. 2008).

    For example, HER2, human epidermal growth factor receptor-2, is a potential target as a diagnostic biomarker. The overexpression of HER2 protein has been observed on the plasma membrane of tumours, in particular breast and ovarian cancers, which is related with poor prognosis (Slamon et al. 1989; Ross & Fletcher 1999). Fluorescent NPs are conjugated with the intact or derived forms of mAb directed against the extracellular domain of HER2 and used to label cancer cells, offering a potential strategy for HER2-targeted diagnostic imaging. It is reported that poly(dl-lactic acid) NPs (PLA NPs) conjugated with anti-HER2 mAb (trastuzumab, Herceptin) can specifically target SKOV-3 human ovarian cancer cells (overexpressing HER2; Cirstoiu-Hapca et al. 2006). A green fluorescent dye, dioctadecyloxacarbo-cyanine perchlorate (Dio), was incorporated into PLA NPs as a fluorescent probe. The internalization of anti-HER2 NPs into SKOV-3 cells was observed in the fluorescent image, representing the efficacy of NPs in active targeting for cancer therapy. Similarly, using the HER2 receptor-specific antibody trastuzumab conjugated to the surface of human serum albumin (HSA) NPs, a specific targeting to HER2-overexpressing cells was reported (Wartlick et al. 2004). Recently, HSA NPs conjugated with trastuzumab and loaded with doxorubicin (DOX) drug were developed (Anhorn et al. 2008). SK-BR-3 breast cancer cells overexpressing HER2 showed a good cellular binding and uptake of the NPs, as well as a specific and efficient growth inhibition after the intake of NPs.

    In addition, QDs are being intensely studied as a class of NP probes for cellular imaging. Wu et al. (2003) have used QDs linked to immunoglobulin G (IgG) and streptavidin to label the breast cancer marker HER2 on the surface of fixed and live cancer cells, to stain actin and microtubule fibres in the cytoplasm, and to detect nuclear antigens inside the nucleus (Wu et al. 2003). All labelling signals are sensitive and specific for the intended targets at the subcellular level. Using QDs with different emission spectra, QD535 (QDs with emission maximum at 535 nm) and QD630, conjugated to IgG and streptavidin, respectively, HER2 receptor and nuclear antigens in SK-BR-3 cells were simultaneously detected with one excitation wavelength. Therefore, QDs conjugated to different targeting moieties are effective in multiplexing assay. Recent advances in molecular, biological and genetic diagnostic techniques have revealed that cancer is controlled by complex multi-functional mechanisms rather than a single factor. The development of fluorescent NPs may contribute significantly to simultaneous and accurate quantification of several cancer-associated biomarkers on a single cell or a small tumour specimen (Jaiswal et al. 2003; Xing et al. 2007; Yezhelyev et al. 2007; Li et al. 2008). For example, Yezhelyev et al. (2007) demonstrated the use of multi-colour QDs for quantitative and simultaneous profiling of multiple biomarkers using FFPE (formalin-fixed and paraffin-embedded) breast cancer cells and FFPE clinical tissue specimens. QDs emitting at 525, 565, 605, 655 and 705 nm were directly conjugated to primary Abs against nuclear hormone receptors (ER and PR), cell membrane surface antigens (HER2 and EGFR) and cytoplasmic mTOR protein. The multi-colour bioconjugates were used for simultaneous detection of the five clinically significant tumour markers in breast cancer cells, MCF-7 and BT-474 (figure 1). Simultaneous quantification of ER, PR and Her2 receptors correlated closely with the results from traditional methods including immunohistochemistry, western blotting and fluorescence in situ hybridization, suggesting that the QD-based technology is well suited for molecular profiling of tumour biomarkers in vitro. Similarly, Fountaine et al. (2006) successfully stained a variety of differentially expressed antigens in FFPE tonsil tissues with up to five different streptavidin-conjugated QDs simultaneously.

    Figure 1.

    Figure 1. MCF-7 and BT-474 human breast cancer cells stained with QD-Abs against nuclear hormone receptors (ER and PR), cell membrane surface antigens (HER2 and EGFR) and cytoplasmic mTOR protein. (a) BT-474 cells showed positive labelling of membrane, cytoplasmic and nuclear antigens indicating expression of all five biomarkers. (b) Spectrum of QD emission of a representative BT-474 cell with emission peaks at 525, 565, 605, 655 and 705 nm, confirming the differential expression levels of the tumour biomarkers (not necessarily the same as the average expression levels of a large cell population). (c) MCF-7 cells labelled with the same panel of QD-Ab bioconjugates. Spectral deconvolution reveals positive stains of ER, EGFR and HER2. (d) Spectrum of QD emission of a representative MCF-7 cell with intensive fluorescent emission at 525 and 655 nm and low signals at 565, 605 and 705 nm (not necessarily the same as the average expression levels of a large cell population). Owing to the nature of spectral deconvolution imaging, (a) and (c) are pseudo-colour images. Scale bar, 10 µm. Adapted from Yezhelyev et al. (2007). Copyright © Wiley-VCH Verlag GmbH & Co. KGaA.

    Owing to their novel optical properties, Yb/Er-doped NaYF4 UCNs were used for cellular imaging of cancer. UCNs were conjugated with anti-HER2 antibody and were used to fluorescently label HER2 receptors of SK-BR-3 cells (Jiang et al. 2009). The intracellular uptake of UCNs via HER2 receptors was visualized using a confocal fluorescence microscope equipped with an NIR laser. The targeted cellular imaging of UCNs provides a potential for in vivo tumour studies by UCNs with excellent optical properties.

    4. In vivo detection of cancer

    In comparison to the study of living cells in culture, different challenges arise with the increase in complexity and size of a multicellular organism. Conventional in vivo imaging methods such as computed tomography (CT) and magnetic resonance (MR) are suitable to delineate morphological features of the tumour, tissue and organs, including the anatomic location, extent and size of the tumour at the macroscopic level (Forstner et al. 1995). Despite continuous improvements in spatial resolution with advanced imaging equipment, CT and MR imaging (MRI) have limited sensitivity and ability to provide specific and functional information on the tumour at the microscopic level. A new field of molecular imaging has been developed for better tumour imaging in living systems (Shah et al. 2004; Atri 2006; Weissleder 2006). Several molecular imaging techniques such as positron emission tomography (PET), single photon emission CT and optical imaging have shown great promises in non-invasive in vivo imaging (Bhushan et al. 2008; Perk et al. 2008; Shi et al. 2008). Especially, fluorescent NPs introduce the possibility of vastly improving sensitivity, resolution and information content of in vivo imaging.

    To develop NPs for promising tumour imaging and eventually translate it to clinical applications, the following issues should be considered. Firstly, NP imaging probes should emit a strong fluorescence signal to improve the detection sensitivity. NPs with NIR excitation (650–900 nm) are highly preferable for in vivo imaging because of their higher penetration depth and minimized tissue autofluorescence compared with UV or visible light (Shah et al. 2001; Vogel & Venugopalan 2003). Secondly, NPs should be photostable, allowing real-time and long-time monitoring of cancer progression during cancer therapy. Thirdly, NPs should be modified with stable and high-affinity targeting moieties, achieving specific targeting of tumour tissues and effective loading to cancer cells. Finally, NPs should be appropriately surface modified, to increase the stability and half-life of NPs in circulation.

    4.1. Targeting of tumours using nanoparticles

    Most anti-cancer agents cannot greatly differentiate between cancerous and normal cells, leading to systemic toxicity and adverse effects. The severe side effects in other tissues greatly limit the maximal allowable dose of the drugs to be systemically delivered in a living system, resulting in inadequate drug concentrations reaching the tumour. NP systems can deliver anti-cancer agents to tumour sites by either passive or active targeting strategy, offering significant benefits to cancer patients (Byrne et al. 2008; Wang et al. 2008).

    Passive targeting takes advantages of the inherent size of NPs and the unique properties of tumour vasculature to enhance the efficacy of drugs. In order to grow beyond 1–2 mm diameter, solid tumours need to increase their surrounding vasculature, in a process known as angiogenesis (Folkman 1990; Folkman & Shing 1992). Angiogenic blood vessels show several abnormalities including a deficiency in pericytes, aberrant basement membrane formation and a relatively high proportion of proliferating cells (Baban & Seymour 1998). The abnormal tumour vasculature results in leaky vessels with gap sizes of 200 nm to 1.2 µm between adjacent endothelial cells (Hobbs et al. 1998; Allen & Cullis 2004), allowing the extravasation of NPs through these gaps into extravascular spaces. In addition, NPs that gain interstitial access to the tumour have higher retention times than normal tissues because of lack of an effective lymphatic drainage that causes an outward convective interstitial fluid flow in tumours (Baish et al. 1996; Baban & Seymour 1998). The leaky vasculature coupled with poor lymphatic drainage induces the enhanced permeability and retention (EPR) effect, resulting in the accumulation of NPs at the tumour site (Jain 2000; Duncan 2003; Brannon-Peppas & Blanchette 2004). There are numbers of studies that display significantly improved therapeutic efficacy of NP drug carriers against different tumour models compared to the free drugs (Nakanishi et al. 2000; Satchi-Fainaro et al. 2004; Okuda et al. 2006; Kim et al. 2008d). The factors that effect the accumulation of NPs in tumours consist of the size and surface characteristics of NPs, and degree of angiogenesis of the tumour, but it is not well understood yet.

    Active targeting is to conjugate targeting moieties to NPs, achieving accumulation of NPs in the tumour sites or individual cancer cells. The targeting moiety, an antibody or ligand, specifically binds to an antigen or receptor overexpressed on the tumour cell surface and assists the NP drug delivery system to selectively and efficiently deliver drugs to tumour sites. The targeted NPs may contribute to the next generation of drug delivery system owing to their increased therapeutic effect (Ulbrich et al. 2004; Xu et al. 2005; Cheng et al. 2007; Diez et al. 2009). However, the biodistribution and pharmacokinetics of the uptake of targeted NPs have not yet been well addressed (Pirollo & Chang 2008). Several recent papers have suggested that antibody targeting primarily increases intracellular uptake of NPs and does not increase tumour localization (Kirpotin et al. 2006; Bartlett et al. 2007; Hussain et al. 2007), while other reports indicate that tumour localization is dependent on the antibody targeting (Wu et al. 2000; Sundaresan et al. 2003). Further studies are clearly needed to understand the systemic delivery of targeted NPs more completely.

    4.2. In vivo tumour imaging

    In vivo diagnostics provide instantaneous data from the patient and allow early detection of cancer and observation of cancer therapy. NPs offer a possibility to produce new medical imaging techniques with higher sensitivity and precision of recognition. Recently, various fluorescent NPs have been used for several types of in vivo animal imaging, including imaging of tumour tissues (Chen et al. 2007; Tanisaka et al. 2008; Miki et al. 2009), angiogenic vasculature (Cai et al. 2006; Smith et al. 2008) and sentinel lymph nodes (SLNs; Kim et al. 2004; Soltesz et al. 2004; Zimmer et al. 2006). Of the various fluorescent NPs investigated to date, NIR fluorescent NPs are of interest for non-invasive in vivo imaging. Organic dye-doped NPs (Chen et al. 2007; Tanisaka et al. 2008; Miki et al. 2009) and QDs (Kim et al. 2004; Soltesz et al. 2004; Cai et al. 2006; Zimmer et al. 2006; Smith et al. 2008) emitting NIR fluorescence (NIRF) have been widely studied in the last decade. Furthermore, UCNs demonstrated their potential in optical imaging of living animals (Chatterjee et al. 2008; Jalil & Zhang 2008).

    Tanisaka et al. (2008) reported that genuine peptide vesicles, peptosomes, composed of hydrophilic poly(sarcosine) and hydrophobic poly(γ-methyl l-glutamate) were labelled with an NIRF probe, indocyanine green (ICG). NIRF imaging of a small cancer on mouse by using peptosome as a nanocarrier was successful due to the EPR effect of the peptosome. Similarly, Miki et al. (2009) demonstrated that ICG labelled amphiphilic block copolymers could be efficiently accumulated in tumours over two weeks via optical imaging of live animals. However, NIR QD spectroscopic properties can be exploited here to achieve deeper penetration than the available NIR dyes (Xing & Rao 2008). Kim et al. (2004) synthesized NIR type II QDs and applied them to SLN mapping in cancer surgery of animals. Injection of only 400 pmol of NIR QD permits SLNs 1 cm deep to be imaged easily in real time using excitation fluorescence rates of only 5 mW cm−2. The utilization of NIR QDs provides the surgeon with direct visual guidance throughout the entire SLN mapping procedure, minimizes incision and dissection inaccuracies and permits real-time confirmation of complete resection. Two years later, the same group reported the synthesis of a size series of NIR QDs smaller than 10 nm at hydrodynamic diameters and their utility in vivo by imaging multiple, sequential lymph nodes and showing extravasation from the vasculature in rat models (Zimmer et al. 2006). Furthermore, Cai et al. (2006) reported the in vivo targeting and imaging of tumour vasculature using arginine–glycine–aspartic acid (RGD) peptide-labelled NIR QDs. QD705 (emission maximum at 705 nm) modified with RGD peptide that specifically targets integrin αvβ3-positive tumour vasculature was intravenously administered into mice bearing subcutaneous U87MG human glioblastoma tumours. The tumour fluorescence intensity reached maximum at 6 h post-injection with good contrast, opening up new perspectives for integrin-targeted NIR optical imaging and imaging-guided surgery. More recently, the same group exploited intravital microscopy with subcellular (approx. 0.5 µm) resolution to examine the binding of neovascularly targeted NIR QDs to tumour vasculature via direct cellular-level visualization in living mice (Smith et al. 2008).

    In addition, while upconversion technology has been known for several decades, the use of UCNs in biology is a relatively recent phenomenon. Zhang's group report that PEI coated NaYF4:Yb,Er UCNs are stable in physiological-buffered saline, non-toxic to bone marrow stem cells and resistant to photo-bleaching (Chatterjee et al. 2008). The UCNs and QDs are injected intradermally and intramuscularly into some tissues either near the body surface or deep in the body of rats. UCNs showed visible fluorescence exposed to an NIR laser, but QDs do not emit fluorescence exposed to UV light (figure 2). Similarly, silica-coated NaYF4:Yb,Er UCNs with strong NIR-to-visible fluorescence were injected intravenously to investigate their biocompatibility and tissue distribution (Jalil & Zhang 2008). The results revealed that UCNs displayed good in vitro and in vivo biocompatibility, demonstrating their potential applications in both cellular and animal imaging systems.

    Figure 2.

    Figure 2. In vivo imaging of rat: QDs injected into translucent skin of (a) foot show fluorescence, but not through thicker skin of (b) back or (c) abdomen; PEI/NaYF4:Yb,Er NPs injected below (d) abdominal skin, (e) thigh muscles or (f) skin of back show luminescence. QDs on a black disc in (a,b) are used as the control. Adapted from Chatterjee et al. (2008). Copyright © 2008 Elsevier.

    4.3. Dual-modality imaging

    The application of optical imaging for in vivo imaging is hampered by limited tissue penetration depth and lack of anatomic resolution and spatial information. Although NIRF imaging can be used to improve penetration depth and detection sensitivity, other imaging modalities, such as MRI and PET, are much better for high spatial resolution, tomographic capability and unlimited tissue penetration depth (Massoud & Gambhir 2003). Thus, a combination of imaging modalities that could allow one to extract anatomical, physiological and molecular information would be advantageous for in vivo imaging (Moore et al. 2004; Medarova et al. 2006; Cai et al. 2007; Lee et al. 2007; Simberg et al. 2007; Yang et al. 2007; Chen et al. 2008; Medarova et al. 2009).

    For example, Moore et al. (2004) synthesized and tested specific and selective accumulation of a novel dual-modal imaging probe in underglycosylated mucin-1 (uMCU-1) antigen-positive subcutaneous tumour in vivo. The dual-modal probe consists of cross-linked iron oxide (CLIO) NP as an MRI contrast agent, modified with Cy5.5 dye as NIRF optical probe, and carrying uMUC-1 targeting peptides attached to the dextran coat of NP. Accumulation of the probe in the tumours after intravenous injection was detected by the decrease in signal intensity on T2-weighted MR images and by the bright NIRF signal on optical images. In 2006, the same group demonstrated that the dual-modal imaging probe could not only detect orthotopically implanted preclinical models of adenocarcinomas but could also track tumour response to 5-fluorouracil, a clinically used chemotherapeutic agent, in vivo in real time (Medarova et al. 2006). More recently, they further proved that tumour delta-T2 was significantly reduced after treatment with DOX drug, indicating a lower accumulation of probe and reflecting the reduced expression of uMUC-1 in orthotopic human breast tumour (Medarova et al. 2009). This dual-modality approach delivers information not only about change in tumour size but also about target antigen expression in vivo. In addition, Cai et al. (2007) and Chen et al. (2008) developed a QD-based dual-function PET/NIRF probe that allows the accurate assessment of the tumour-targeting efficacy of QDs. Yang et al. (2007) report the synthesis of core-cross-linked polymeric micelles (CCPMs) entrapped with Cy7 and labelled with 111In, and significant accumulation of CCPMs in human breast tumour xenografts via both NIRF optical and gamma imaging techniques.

    5. Multi-functional nanoparticles

    NPs have a large surface area to accommodate a large number or a wide range of surface functional groups allowing chemical conjugation to multiple diagnostic and therapeutic agents as well as targeting moieties. This may enable the development of multi-functional ‘smart’ NPs for simultaneous tumour imaging and treatment, a major goal in cancer research. For example, a core particle could be conjugated to a specific targeting component that provides preferential accumulation of NPs in the target tumour. Simultaneously, the same particles could be linked to a therapeutic agent to inhibit the tumour growth and an imaging agent to monitor the drug transport process (figure 3). Although these new materials are of great interest, most of the studies are still at an early or proof-of-concept stage with cultured cancer cells and in vivo imaging and are not relevant to the treatment of solid tumours.

    Figure 3.

    Figure 3. Multi-functional NP. The multifunctional NP has the capability to simultaneously carry therapeutic agents (squares), imaging contrast agents (diamonds) and targeting moieties (circles). The NP can be used for specific delivery of anti-cancer agents, tracking of therapeutic delivery and detection of treatment effects in real time.

    Fluorescent NPs have shown a great promise for cellular labelling and in vivo imaging of cancer. Monitoring of the biodistribution of particles can be done in real time. Drug-loaded NPs also have been used for treating cancer in animal models and even in clinical trials. Owing to their size and structural similarities, it is possible that they can be integrated and multi-functional NPs can be constructed to simultaneously track therapeutic delivery and treat cancer.

    Multi-functional NPs are prepared with organic materials and/or inorganic materials. Organic liposomes, polymer NPs and carbon nanotubes have been used as drug carriers that are loaded with therapeutic agents (anti-cancer drugs, protein, nucleic acid, photothermal agents, etc.) and modified with target moieties and imaging agents (magnetic NPs, organic dyes, QDs, gold NPs, etc.; Huang et al. 2007; Guo et al. 2008; Kim et al. 2008c; Weng et al. 2008). On the other hand, inorganic NPs, such as QDs, magnetic NPs and lanthanide-doped UCNs, can be used as a central core that defines their fluorescent, magnetic and optical properties, and then modified with polymer or silica coating in order to improve their biostability and biocompatibility, and finally loaded with therapeutic agents and conjugated with targeting moieties (Bagalkot et al. 2007; Kim et al. 2008b; Yezhelyev et al. 2008).

    NPs are known for their versatility, and various multi-functional NPs capable of simultaneous imaging, delivery, etc. are used. For example, a multi-functional micelle was developed for in vitro cancer cell targeting, distribution imaging and anti-cancer drug delivery (Huang et al. 2007). The mixed micelles were prepared by dialysis from graft copolymer and diblock copolymer, and then DOX, an anti-cancer drug, was incorporated into the inner core of micelles. FITC-labelled micelles exhibited a specific tumour targeting mediated by asialoglycoprotein receptors and a clear distribution in the cytoplasm. DOX drugs released from micelles have strong effects on the viability of tumour cells HepG2. Another polymeric NP, PLGA, was used as a matrix for loading magnetite nanocrystals or QDs, in addition to DOX drugs (Kim et al. 2008c). Pegylated folate was coated on the surface of NPs for active targeting of cancer cells. Cancer cell-targeted MRI and optical imaging, as well as drug delivery, were successfully demonstrated by the multi-functional polymer NPs. Furthermore, several kinds of targeted delivery system were developed for tumour labelling applications in vivo. Liposomes, the most clinically established NPs for drug delivery, were conjugated to luminescent QDs and tumour-targeting moieties (e.g. anti-Her2 single chain Fv fragments) and were loaded with DOX (Weng et al. 2008). DOX-loaded NPs showed efficient anti-cancer activity in HER2-overexpressing SK-BR-3 cells. The QD-conjugated immunoliposome-based NPs were able to provide tumour cell imaging and in vivo localization at tumour sites after their intravenous injection. Owing to the superior optical properties (e.g. brightness and photostability) of QDs, they were also functionalized to the outer surface of carbon nanotubes for in vivo trackable delivery. Shi et al. (2007a) synthesized CdSe/ZnS QD-conjugated carbon nanotubes (CNT-QD) by a unique plasma coating method and demonstrated their potential applications in cancer diagnosis. One year later, the same group reported that CNT-QD exhibited strong luminescence suitable for non-invasive optical in vivo imaging with an NIR emission around 800 nm (Guo et al. 2008). Paclitaxel, an anti-tumour agent, loaded CNT-QD demonstrated in vitro anti-tumour efficacy against human PC-3MM2 prostate cancer cells. However, the nanotubes will not be used for in vivo tumour treatment unless they can lead paclitaxel to tumour specifically after intravenous injection.

    Among these various inorganic nanomaterials, silica or mesoporous silica materials have been intensively investigated for their potential application as delivery vehicles (Lai et al. 2003; Lu et al. 2007; Slowing et al. 2007) and imaging probes (Huang et al. 2005; Yi et al. 2005; Lin et al. 2006; Selvan et al. 2007; Wu et al. 2008; Lee et al. 2009) owing to their considerable biocompatibility and easy surface functionalization. However, these materials have not been used for in vivo simultaneous imaging and drug delivery. Kim et al. (2008b) presented discrete, monodisperse and precisely size-controllable core–shell mesoporous silica NPs smaller than 100 nm. A single Fe3O4 nanocrystal core was used as an MRI agent, and FITC or RITC incorporated covalently into the silica wall was used as a fluorescent imaging agent. NPs were verified to accumulate preferentially at tumour sites through the EPR effect after intravenous injection. To examine the drug delivery of the NPs, the cytotoxic effect of DOX-loaded NPs was tested on SK-BR-3 cells, indicating that mesoporous silica NPs have a potential for drug delivery into cancer cells to induce cell death. In another example, Bagalkot et al. (2007) reported a novel QD-aptamer-DOX conjugate (QD-Apt(DOX)) as a targeted cancer imaging, therapy and sensing system. By functionalizing the surface of fluorescent QD with the A10 RNA aptamer, which recognizes the prostate-specific membrane antigen, and intercalating DOX into the double-stranded CG sequence of the aptamer, a targeted QD-Apt(DOX) conjugate with reversible self-quenching properties based on a Bi-FRET mechanism was developed. Two donor–quencher pairs of FRET occurred in this construct, as the QD fluorescence was quenched by DOX, and DOX was quenched by aptamers. The multi-functional NP system can deliver DOX to the targeted prostate cancer cells and sense the release of DOX by activating the fluorescence of the QD, which concurrently images the cancer cells and inhibits the growth of specific prostate cancer cell lines. However, the current design is not sufficient for in vivo use unless the drug-loading capacity can be greatly increased.

    Multi-functional NPs can also be used for small interference RNA (siRNA) delivery and imaging. RNA interference (RNAi) has emerged as a powerful technology for sequence-specific silencing of genes and holds great potential as a new treatment paradigm for human diseases amenable to manipulation at the gene expression level (Fire et al. 1998; Elbashir et al. 2001; Bumcrot et al. 2006; Rana 2007; Castanotto & Rossi 2009; Kurreck 2009). The basic mechanism of RNAi is thought to be a multi-step process where siRNA (usually double-strand RNA 21–23 nucleotides in length) with sequences complementary to a gene's sequence suppresses the activity of the gene by bringing about the degradation of the corresponding messenger RNA (mRNA), thus blocking the translation of the mRNA into proteins (Fire et al. 1998; Elbashir et al. 2001; Hannon 2002; Martinez et al. 2002). The major obstacle to therapeutic application of RNAi, however, is the delivery of siRNA to the target tissue owing to its fast degradation in the physiological environment, poor cellular uptake, inefficient endosomal escape, insufficient dissociation from the carrier and lack of targeting ability. Numerous siRNA delivery carriers have been developed such as liposomes (Hassani et al. 2005; Landen et al. 2005; Sonoke et al. 2008; Villares et al. 2008), polymers (Takeshita et al. 2005; Urban-Klein et al. 2005; Xiong et al. 2009), inorganic NPs (Liu et al. 2007b; Medarova et al. 2007), aptamer–siRNA chimaeras (McNamara et al. 2006), cholesterol-conjugated siRNAs (Soutschek et al. 2004) and protamine–antibody fusion protein (Song et al. 2005). Additionally, there is a need for a method to non-invasively track and monitor siRNA delivery to tissues of interest in order to achieve minimally invasive administration in clinical application. Some typical strategies have been applied to optical tracking of siRNA, involving fluorescently end-modified siRNAs (Howard et al. 2006), bioluminescence imaging of siRNA-mediated silencing (McCaffrey et al. 2002; Muratovska & Eccles 2004; Takeshita et al. 2005) and co-transfecting siRNA with fluorescent NPs (Medarova et al. 2007; Tan et al. 2007; Yezhelyev et al. 2008). Therefore, a multi-functional system is developed to simultaneously deliver and image siRNA for in vitro and in vivo detection.

    For example, Tan et al. (2007) used chitosan NPs with encapsulated QDs and surface-conjugated anti-HER2 antibody for targeted delivery of HER2/neu siRNA into HER2-overexpressing SK-BR-3 breast cancer cells. Using such a construct, the intracellular delivery of siRNA was monitored by the presence of fluorescent QDs in the chitosan NPs and gene silencing effects of conjugated siRNA were also established using the luciferase and HER2 ELISA assay. Similarly, Yezhelyev et al. (2008) reported that QDs were used as a nanometre-sized scaffold to conjugate siRNA for real-time tracking and 10- to 20-fold improvement in gene-silencing efficiency. The main finding is that a proton-sponge layer formed on the QD surface by balancing two function groups (carboxylic acid and tertiary amine) leads to efficient siRNA release from acidic endosomes and into cytoplasm. As a result, gene silencing activity is improved by 10- to 20-fold and cell toxicity is reduced by 5- to 6-fold in MDA-MB-231 cells (in comparison with the current siRNA delivery agents). As another example, silica-coated NaYF4:Yb,Er UCNs can serve as a multi-functional system for simultaneous imaging and delivery of siRNA in vitro (Jiang et al. 2009). siRNA was attached to anti-Her2 antibody-conjugated UCNs and the delivery of these NPs to SK-BR-3 cells overexpressing HER2 receptors was visualized using confocal microscopy. Meanwhile, a luciferase assay was established to confirm the gene-silencing effect of delivered siRNA. However, they have only shown applicability in cell culture, but not yet in living animals. For in vivo cancer therapy, Medarova et al. (2007) described the synthesis and characterization of multi-functional NPs for in vivo transfer of siRNA and the simultaneous imaging of its accumulation in tumours by high-resolution MRI and NIR in vivo optical imaging (NIRF). The NPs used CLIO NPs as a core and MRI contrast agent, Cy5.5 dye labelled on the surface as an NIRF optical agent and siRNA duplex conjugated to target a gene of interest. Myristoylated polyarginine peptides were also conjugated on the surface serving as a membrane translocation module. In a series of experiments targeting model (green fluorescence protein) and therapeutic genes (survivin), they found that the delivery system could be monitored in vivo by MRI and optical imaging. In addition, the silencing process of delivered siRNA was able to be corroborated by proof-of-principle optical imaging and further confirmed by histological data (figure 4). Overall, the new approach could advance the development and optimization of cancer therapeutic product, siRNA. However, the ‘smart’ multi-functional NPs are still in development and have only recently shown applicability in imaging in living animals, but not yet in in vivo tumour therapy.

    Figure 4.

    Figure 4. In vivo imaging of MN-NIRF-siGFP silencing in tumours. (a) In vivo NIRF optical imaging of mice bearing bilateral 9L-GFP and 9L-RFP (GFP, green fluorescent protein; RFP, red fluorescent protein) tumours 48 h after intravenous probe injection. There was a marked decrease in 9L-GFP-associated fluorescence (P¼ 0.0083) and no change in 9L-RFP fluorescence. To generate GFP/RFP reconstructions, GFP and RFP images were acquired separately and then merged. (b) Correlative ex vivo fluorescence optical imaging showed a significant drop in fluorescence intensity in 9L-GFP tumours (P¼ 0.0036). There was no evidence of silencing in saline-injected controls. (c) Confocal microscopy of frozen tumour sections indicated the presence of the probe in both 9L-GFP and 9L-RFP tumours. However, in GFP tumours, GFP fluorescence was predominantly at background levels. Note that the 9L-RFP section produced bright fluorescence. Scale bar, 20 mm. (d) Quantitative RT–PCR analysis of GFP expression performed on total RNA extracted from 9L-GFP tumours from mice injected with either MN-NIRF-siGFP, a mismatch control or saline solution. Representative data are shown. Adapted from Medarova et al. (2007). Copyright © 2007 Macmillan Publishers Ltd.

    6. Conclusion

    Although nanotechnology is still at an early stage of development, the possibilities for new therapies using this technology to treat cancer seem to be very numerous. Fluorescent NPs appear very promising as novel tools for highly sensitive optical imaging of cancer at cellular and animal levels. Furthermore, the emerging development in novel multi-functional NPs, which combine different functions in one particle, provides new potential for clinical therapies and diagnostics and undoubtedly will make an impact on cancer therapy. With this promising progress in the development of fluorescent NP-guided cancer therapy, it is imperative to constantly improve the fundamental understanding of designing and applying NPs for diagnosis, treatment or the combination of imaging and therapeutics in different clinical situations. However, our limited knowledge on the toxicity of these NPs limits us from exploiting their full potential in clinical applications. Their potential toxicity and their possible incompatibility may result in disorders such as inflammation, immunoreaction or even cancer. The mechanisms of those effects are not well studied yet, but should depend more on whether their accumulation in specific organs occurs because such a deposition of NPs could provoke intracellular changes that might affect cell integrity and hence organ function. Thus, more studies are needed to explore these aspects and new methods that improve the biocompatibility of NPs should be developed to reduce the undesirable side effects. Once these questions have been addressed, optimized NPs with improved selectivity, efficiency and safety could be designed and applied for the diagnosis and treatment of cancer.

    Acknowledgements

    The authors acknowledge financial support from A*STAR BMRC (R-397-000-062-305) and the National University of Singapore.

    Footnotes

    © 2009 The Royal Society

    References

    Previous Article
    Next Article