Intracellular uptake, transport, and processing of gold nanostructures

Abstract

The emerging field of nanomedicine requires better understanding of the interface between nanotechnology and medicine. Better knowledge of the nano-bio interface will lead to better tools for diagnostic imaging and therapy. In this review, recent progress in understanding of how size, shape, and surface properties of nanoparticles (NPs) affect intracellular fate of NPs is discussed. Gold nanostructures are used as a model system in this regard since their physical and chemical properties can be easily manipulated. The NP-uptake is dependent on the physiochemical properties, and once in the cell, most of the NPs are trafficked via an endo-lysosomal path followed by a receptor-mediated endocytosis process at the cell membrane. Within the size range of 2–100 nm, Gold nanoparticles (GNPs) of diameter 50 nm demonstrate the highest uptake. Cellular uptake studies of gold nanorods (GNRs) show that there is a decrease in uptake as the aspect ratio of GNRs increases. Theoretical models support the size- and shape-dependent NP-uptake. The intracellular transport of targeted NPs is faster than untargeted NPs. The surface ligand and charge of NPs play a bigger role in their uptake, transport, and organelle distribution. Exocytosis of NPs is dependent on size and shape as well; however, the trend is different compared to endocytosis. GNPs are now being incorporated into polymer and lipid based NPs to build multifunctional devices. A multifunctional platform based on gold nanostructures, with multimodal imaging, targeting, and therapeutics; hold the possibility of promising directions in medical research.

Keywords::

• Gold nanoparticles
• Gold nanorods
• uptake
• transport
• organelle distribution

Introduction

The field of nanomedicine, the use of the tools and knowledge of nanotechnology for biomedical purposes, is currently undergoing an explosive development on many fronts (Alivisatos 2003). Recent advances in engineering and technology have led to the development of many new nanoscale biomedical platforms, including quantum dots, nanoshells, GNPs, paramagnetic NPs, carbon nanotubes, and improvements in traditional, lipid-based nanoscale platforms. In particular, GNPs have been explored as a model platform for biomedical research due to their favourable physical and chemical properties (Bergen et al. 2006). The size and shape of GNPs can be tailored to range from 2–100 nm and their surface properties allow for facile functionalization and targeting to specific biological structures such as the nucleus (Feldherr et al. 1984, Tkachenko et al. 2004, Souza et al. 2006, Berry et al. 2007, Oyelere et al. 2007, Jiang et al. 2008, Nativo et al. 2008). In addition, the ability to produce various delivery forms like liposomes, micelles or dendrimers has increased the application scope of GNPs (Carrot et al. 1998, Mohamed et al. 1998, Garcia et al. 1999, Sung-Hee et al. 2006). These advantages, along with their biocompatibility, have motivated interest in employing gold nanostructures in cell imaging, targeted drug and gene delivery, and biosensing (Sandhu et al. 2002, Sokolov et al. 2003, Shukla et al. 2005, Bergen et al. 2006, Han et al. 2006, Kneipp et al. 2006, Kumar et al. 2007). Also, GNPs have been receiving significant attention for use in cancer diagnosis and treatment (Hainfeld et al. 2004, Loo et al. 2005, El-Sayed et al. 2006, Huang et al. 2006, Wijaya et al. 2009, Zhang et al. 2009a, Zheng & Sanche, 2009, Brown et al. 2010, Chithrani et al. 2010). There have been a number of studies investigating the potential cytotoxic effects, and intracellular behaviour of GNPs as a function of their physiochemical properties (Connor et al. 2005, Shukla et al. 2005, Chithrani et al. 2006, Jiang et al. 2008, Lewinski et al. 2008). Most of these investigations have used static methods such as inductively coupled surface plasmon atomic emission spectroscopy, transmission electron microscopy (TEM), and fixed-cell confocal microscopy. However, understanding of the cytoplasmic transport of gold nanostructures in four dimensions (space and time-4D) provides new insight into their intracellular behaviour. As a first step in this direction, surface-enhanced raman spectroscopy (SERS) and confocal microscopy have been used to probe the interactions of NPs with the cellular environment (Kneipp et al. 2006, Huff et al. 2007, Kumar et al. 2007, Chithrani et al. 2009a, Chithrani et al. 2009b).

In this review, recent progress in GNP-based research work towards understanding of how physiochemical properties of gold nanostructures (throughout the article, spherical colloidal gold nanostructures are referred to as GNPs while rod-shaped gold nanostructures are referred to as GNRs) affect their intracellular fate will be discussed. Specifically, this review is focused on summarizing answers to the following fundamental questions:

• (1) What is the dependence of the endocytosis and exocytosis process on size, shape and surface properties of NPs?

• (2) What are the transport properties of these NPs?

• (3) What is the progress on nuclear targeting of these NPs?

• (4) What are the applications of these NPs?

• (5) How do we incorporate these NPs to existing delivery systems to built better multifunctional devices for therapy and imaging in medical research?

Figure 1A is a schematic diagram that highlights some of the important cellular processes involving GNPs that will be discussed in this review. As illustrated in Figure 1, the NPs are first internalized by cells through receptor-mediated endocytosis (RME) and trapped in organelles called ‘endosomes’ (Chithrani & Chan 2007). These endosomes then fuse with lysosomes for processing before being transported to the cell periphery for excretion. These different stages of NP transport through the cell captured by TEM are illustrated in Figure 1B. In the first part of the review, the current knowledge about how physiochemical properties affect cellular uptake of NPs is discussed. In the second part, we will discuss current understanding of the transport properties, organelle distribution, and exocytosis of NPs. In the third part, nuclear targeting of GNPs and their applications will be discussed. Finally, the feasibility of incorporating gold nanostructure into future generations of multifunctional NPs will be discussed for applications in cancer therapy and imaging.

Figure 1. Intracellular uptake, transport, processing, and excretion of NPs. (A) Schematic describing endo-lysosomal pathway (left) and RME process (right) of NPs inside the cell. NPs are internalized by receptor-mediated endocytosis and trapped in endosomes. These endosomes fuse with acidic organelles, lysosomes, for processing. Finally they are transported to the cell periphery for excretion. (B) TEM images capturing different stages of NP transport through the cell. Reproduced with permission from [Chithrani et al. 2006, Chithrani & Chan 2007, Jin et al. 2009].

Cellular uptake of NPs

Mechanism of NP cell uptake

Endocytosis is one of the major pathways for cellular uptake of NPs (Kam et al. 2004, Chithrani et al. 2006, Chithrani et al. 2009b). Particularly for most gold nanostructures, the internalization mechanism is considered to be receptor-mediated endocytosis (RME) (Mukherjee et al. 1997, Kirchhausen 2000, Chithrani et al. 2006, Kam et al. 2006, Jin et al. 2009). RME of NPs occurs through interactions between proteins (ligands) on the surface of NPs and cell membrane receptors (inset Figure 1A). The ligand-receptor relationship has been exploited in applications to improve targeting of NPs (Supplementary Section S1 – online only). However, for uncoated NPs, RME arises from the adhesion of media proteins to the surfaces of NPs during a typical cell culture incubation experiment (Chithrani et al. 2006, Lynch 2007). The cell surface receptors binding to the curved NP surface causes membrane curvature with a corresponding increase in elastic energy. This receptor-ligand binding also causes configurational entropy to be reduced due to immobilization of the receptors and at the same time, receptors can diffuse to the wrapping site driven by the local reduction in free energy, allowing the complete membrane wrapping around the particle (Jin et al. 2008). Hence, RME of NPs is an energy dependent process. The uptake of NPs is decreased at low temperature (4°C) and in ATP (adenosine try phosphate)-depleted environments (cells pretreated with NaN₃) indicating that gold nanostructures enter cells via RME (Silverstein et al. 1977, Schmid & Carter 1990, Kam et al. 2006, Chithrani & Chan 2007). After internalization through RME, these NPs localize in endosomes before being fused with lysosomes for degradation (Figure 1). However, a different uptake mechanism is suggested for ligand-free GNPs as discussed later in section 4.0 (Taylor et al. 2010).

Dependence of cell uptake on size of NPs

Recent studies have identified that the size of GNPs plays an important factor in their cellular uptake process (Xu et al. 2004, Chithrani et al. 2006, Arnida & Ghandehari 2009). Stability and uniformity in size distribution of NPs are essential for accurate determination of NP uptake. GNPs were first synthesized in aqueous solution by Faraday (Faraday, 1857). Turkevich et al. and Frens reported growth of GNPs by citrate reduction (Turkevich et al. 1951, Frens 1973). Recently, several methods have been proposed to verify the stability and reduce the polydispersity of NPs (Supplementary Section S2 – online only) (Murray et al. 2000, Lévy et al. 2004, Ji et al. 2007). GNPs with diameter ∼ 50 nm exhibited significantly higher uptake compared to smaller or larger NPs (Figure 2A) (Chithrani et al. 2006, Arnida & Ghandehari, 2009). Aoyama and coworkers have also demonstrated that the RME of NPs is strongly size-dependent and the optimum NP-diameter for uptake is ∼ 50 nm (Aoyama et al. 2003, Nakai et al. 2003, Osaki et al. 2004). Similar size-dependent NP-uptake was seen for other NPs such as Ag (Xu et al. 2004). However, it is important to mention that rate and extent of NP-uptake can be varied among different cell lines (Cartiera et al. 2009).

Figure 2. Dependence of cellular uptake of NPs as a function of size and shape. (A) Variation of cellular uptake of NPs as a function of size. (B) Model explaining the size- and shape-dependent cell uptake of NPs (spherical-shaped NPs = blue; cylindrical-shaped NPs = pink; comparison with the model put forward by Freund and co-wokers (green)). (C) Comparison of GNR uptake with their spherical counterparts (D) Shape-dependent cellular uptake of carbon nanotubes. Reproduced with permission from [Chithrani et al. 2006, Jin et al. 2009].

Theoretical models have been established to provide insights into the dynamics of size- and shape-dependent RME of NPs based on energetic and kinetic considerations (Bao & Bao 2005, Gao et al. 2005, Shi et al. 2008, Zhang et al. 2009b). It has been predicted that GNPs of diameter 50 nm have the shortest internalization time of 20 minutes, which appears to be consistent with experimental data (Aoyama et al. 2003, Nakai et al. 2003, Osaki et al. 2004, Gao et al. 2005). It is interesting to note that there is an optimal particle diameter for the smallest internalization time. The optimal particle size is a result of competition between thermodynamic driving forces and receptor diffusion kinetics. This model elucidates the mechanism of size-dependent cellular uptake of NPs from a kinetic point of view and has addressed ‘how fast’ a single NP can be endocytosed into the cell. However, Suresh and co-workers went one step further and have addressed the question of ‘how many’ NPs can be endocytosed in a sufficiently long period of time using thermodynamic arguments (Zhang et al. 2009b). However, the uptake of smaller NPs (diameter less than 44 nm) cannot be explained using the above mentioned models. Recently, another theoretical model has been put forward by Strano and co-workers to address this issue (Jin et al. 2009). According to their model, a surface clustering on the external cellular membrane facilitates RME by lowering the otherwise prohibitive thermodynamic barrier for smaller NPs. This model offers a plausible explanation of the experimentally observed size-dependent uptake of NPs (Figure 2B).

Dependence of cell uptake on shape of NPs

The cellular uptake of NPs is dependent upon shape as well. The uptake of shorter GNRs is higher than longer nanorods (Figure 2C) (Chithrani et al. 2006). In addition, the uptake of rod-shaped GNPs is lower than their spherical counterparts. One reason could be the difference in the curvature of the different-shaped NPs. For example, the rod-shaped NPs can have larger contact area with the cell membrane receptors than the spherical NPs when the longitudinal axis of the rods interacts with the receptors. This could reduce the number of available receptor sites for binding. A second reason could be the amount of CTAB (Cetyl trimethylammonium bromide) surfactant molecules adsorbed onto the rod-shaped NP surface during synthesis. If the CTAB is still on the surface, the serum protein may not be able to bind onto the GNP surface efficiently. Also, the protein coating on the surface of the rod-shaped gold NPs may not be homogeneous. In such a case, the proteins on the surface of the rod-shaped GNPs may not bind to receptors on the cell surface as strongly.

The cell uptake properties of rod-shaped NPs has been explored using other NP-systems such as carbon nanotubes and the results are consistent with the GNR studies (Figure 2D) (Jin et al. 2009). The experimental data shows that the longer rod-shaped (higher aspect ratio) NPs have a lower uptake as compared to shorter rod-shaped (lower aspect ratio) NPs (Figure 2C,D). This is supported by theoretical model calculations (Shi et al. 2008, Jin et al. 2009). According to the recent model put forward by Gao and co-workers, such shape-dependent uptake is interpreted as a result of balance between the diffusion constant of the particles and the interaction energy between the particles and the cell relative to the thermal energy (Shi et al. 2008). They have used hydrodynamic radius of rod-shaped NPs to explain the shape-dependent uptake. Particles with smaller hydrodynamic radii have larger diffusion constant but weaker interaction with the cell while larger particles have smaller diffusion constant but stronger interaction with the cell (Shi et al. 2008). Strano and co-workers have used one of Gao's previous models for rod-shaped NPs, but introduced an effective scaling metric called ‘capture radius’ to explain the shape-dependent uptake of NPs (Jin et al. 2009). This model can be used to explain the trend in uptake of not only nanotubes, but also GNRs (Figure 2B).

Dependence of cell uptake on surface of NPs

The effect of surface ligands on cell uptake has been explored by coating NPs with different proteins (such as epidermal growth factor (EGF), transferin) and it was found that these protein coated GNPs have similar uptake properties to those coated with serum protein, with a maximum uptake for NPs with diameter 50 nm as previously illustrated in Figure 2A (Chithrani & Chan 2007, Chithrani et al. 2009b). However, the uptake of protein-coated GNPs is less than for serum protein-coated ones. This result is consistent since, for example, tranferrin represents only one kind of protein in cell media and cells display multiple receptor types, diminishing the surface density of tranferrin receptors. As a result, tranferrin receptors are quickly saturated by the transferrin-coated GNPs while for serum-coated GNPs, the surface density of usable receptors is much larger, since many kinds of receptors are available. The dependence of GNR uptake on surface ligand is similar to that of GNPs (Huff et al. 2007). Uptake efficacy of NPs can be enhanced by exploiting the ligand-receptor relationship for cells with particular over expressed receptors (Supplementary Section S1).

Recent studies have shown that the cellular uptake of GNPs was significantly affected by the surface charge of the NPs as well (Cho et al. 2009, Liang et al. 2010). The uptake efficiency of the positively charged NPs was greater than that of the neutral and negatively charged NPs (Figure 3A). The high uptake of positively charged NPs was explained using a theoretical model (Cho et al. 2009). According to this model, the uptake process occurs in two steps: adsorption onto the membrane of the cell and internalization by the cell. As explained in inset figure 3A, positively charged NPs should adhere to the negatively charged cell membranes and facilitate the higher uptake into cells. Interactions with some surface molecules on cell membranes may be responsible for the facilitated uptake of negatively charged NPs. In addition, it has shown that the charge of the NP plays a bigger role in NP-uptake than its size (Liang et al. 2010). Consistent with the previous discussion, several groups have found that positively charged GNRs exhibited higher cellular uptake than negatively charged ones (Hauck et al. 2008, Alkilany et al. 2009).

Figure 3. Dependence of cell uptake on surface charge. (A) Effect of surface charge of GNPs on cell uptake for SK-BR-3 cells. Inset schematics illustrate interactions between cell membrane interactions and GNPs with different surface charges (B) Effect of NP surface charge (top) and size (bottom) of NP on protein structure (top). Reproduced with permission from Vertegel et al. 2004, Aubin-Tam & Hamad-Schifferli 2005, and Cho et al. 2009.

As discussed before, the uptake efficiency of NPs can be tailored by conjugation of specific proteins onto the NP-surface. However, recent studies have shown that the protein conjugation is also affected by the surface charge and size of the NPs (Vertegel et al. 2004, Aubin-Tam & Hamad-Schifferli 2005, Aubin-Tam & Hamad-Schifferli 2008). According to these studies, denaturation of proteins occurs when NPs have either positively or negatively charged ligands, whereas they are not denatured at all when linked to NPs with neutral ligands (Figure 3B) (Aubin-Tam et al. 2008). NP size can also affect the protein structure and activity (Figure 3B). For larger NPs, the effective surface area it can access is larger increasing the likelihood of denaturing the protein. For smaller NPs, less denaturation occurs due to less surface area and fewer numbers of ligands that can interact with the protein (Aubin-Tam & Hamad-Schifferli 2008). Hence, it is important to consider surface charge and size of NPs while designing NP-vectors for different applications.

Based on these fundamental studies in vitro, physiochemical dependent NP-uptake cannot be ignored when designing NP-based systems for biomedical applications. Recent in vivo studies have shown that the accumulation of NPs in a tumour also depends on the size and surface properties of NPs as well. Yang et al. have used polymer NPs of sizes 20, 40, and 60 nm and found that NPs of diameter 60 nm have higher tumour accumulation (Yang et al. 2009). Ishida et al. have used liposomes in the size rage 100–400 nm and shown that liposomes of size 100 nm have a higher tumour accumulation compared to larger sizes (Ishida et al. 1999). However, Chan and co-workers did not observe a size-dependent restriction on sub-100 nm particle extravasations from the blood to the tumour (Perrault et al. 2009). It has been shown recently that targeted GNPs can provide greater intracellular delivery of therapeutic agents to the cancer cells within solid tumours than their untargeted analogs. For example, transferrin-coated GNPs showed greater accumulation in the tumour than the untargeted ones (Choi et al. 2010). However, further studies are required to elucidate the physiochemical-dependent NP-uptake during in vivo experiments in order to design better NP-based carriers for biomedical applications (Rao 2008).

Intracellular transport kinematics of NPs

Particle transport rates in cell cytoplasm are heterogeneous. For example, following cell entry, NPs, non-viral gene carriers, and viral particles are transported at various rates with some particle systems moving several orders of magnitude faster than the others (Supplementary Section S3 – online only) (Suh et al. 2003, Suh et al. 2004, Suh et al. 2005, Suk et al. 2007). Transport modes of all intracellular complexes can be divided into diffusive, subdiffusive, and active transport (Suh et al. 2005). Most of the studies have so far focused on transport of viral and non-viral gene carriers. These gene carriers are actively transported by motor proteins along microtubules (Suh et al. 2003, Suh et al. 2004, Suh et al. 2005, Suk et al. 2007). The average velocity of actively transported gene vectors is 0.2 μm/s, which is of the same order of magnitude as movement along microtubules involving motor proteins such as dyneins and kinesins (King & Schroer 2000, Suh et al. 2003, Suh et al. 2004). However, 85% of the viral and non-viral gene vectors are transported slowly and nonactively through either diffusive or subdiffusive transport modes.

Transport of gold nanostructures has been studied by several research groups (Huff et al. 2007, Jin et al. 2008, Chen & Irudayaraj 2009, Chithrani et al. 2009b). As illustrated in Figure 4, targeted and untargeted NPs seem to display different transport properties. Untargeted gold nanostructures display slower diffusion in the cell cytoplasm. The diffusion coefficients for GNPs and GNRs are 0.001 and 0.0004 μm²/s, respectively (Figure 4A,B). The values corresponding to diffusion coefficient are comparable to diffusive transport properties of viral and nonviral gene carriers (Supplementary Table S3). In contrast, targeted GNRs display active transport behaviour similar to actively transported viral and nonviral gene carriers (Chen & Irudayaraj 2009). The diffusion coefficient for targeted GNRs is 7.7 μm²/s and comparable active transport properties are seen for DNA conjugated nanotubes (Jin et al. 2008). This type of active transport of targeted NPs is believed to be mediated through motor proteins along microtubules (Jin et al. 2008). In addition, studies on transport dynamics of NPs in endosome and lysosomes shows slower diffusion time for NPs localized in lysosomes compared to endosomes as illustrated in Figure 4C. Generally, the size of lysosome is bigger than the endosome because it is formed by fusion of transport vesicles rich in hydrolases. It is also known that the cellular transport is size dependent as well (Lukacs et al. 2000). Hence, the larger size of the lysosome may have led to the slower diffusion of NPs in lysosomes compared to endosomes (Chen & Irudayaraj 2009). It is also known that these organelles move along microtubules and hence, we believe that endosomes and lysosomes carrying these NPs also travel along microtubules similar to the transport of viral and non-viral gene carriers (Cordonnier et al. 2001, Derfus et al. 2004, Chen & Irudayaraj 2009, Kulkarni et al. 2006). However, more studies are needed to elucidate transport dynamics based on size, shape, surface properties, and NP-material. Understanding the transport properties of NPs in cellular environments is crucial to the rational design of multifunctional NPs for targeting and delivery.

Figure 4. Transport properties of GNPs. (A,B) Transport properties of untargeted GNPs and GNRs, respectively (C) Top: Experimental setup used for tracking of transport in organelles; Bottom: comparison of diffusion properties of targeted GNRs in endosomes and lysosomes. Reproduced with permission from Huff et al. 2007, Chen et al. 2009, and Chithrani et al. 2009b.

Intracellular processing and exocytosis of NPs

Intracellular processing of NPs

One of the basic cellular biological processes of NPs is outlined in Figure 1. The nanostructures are first taken up by cells through the RME process and localize in organelles such as endosomes and lysosomes (Goldstein et al. 1979). However, the intracellular processing can be varied for RME of targeted and untargeted NPs as illustrated in Figure 5. The processing of untargeted GNPs is fast and over 75% of the NPs end up in lysosomes within 45 minutes of incubation. However, the dynamics and temporal distribution of targeted GNRs in organelles are different as illustrated in Figure 5B (Chen & Irudayaraj 2009). Herceptin-conjugated GNRs (H-GNRs) show higher accumulation in endosomes than in lysosomes and the concentration of H-GNRs in the endosomes was 7.5 times higher than in the lysosomes. Only a small amount of GNRs was found to escape from the endo-lyso pathway to the cytosol. The intracellular processing is further discussed in the next section to illustrate how surface properties and NP-material can influence the organelle distribution of NPs.

Figure 5. Distribution of GNPs and GNRs in organelles. (A,B) Schematic and experimental data describing the processing of GNPs and GNRs after endocytosis for MCF-7 and SK-BR-3 cells. [Lévy et al. 2004]. Reproduced with permission from Chen & Irudayaraj 2009, and Chithrani et al. 2009b.

It has also been shown that certain biomolecules on the NP surface play a bigger role in their intracellular processing (Sée et al. 2009). GNPs conjugated with certain peptides are degraded within the endosomal compartments through peptide cleavage by the protease cathepsin L. The cathepsin L is able to cleave more than a third of the human proteome, indicating that this degradation process is likely to happen to most NPs conjugated with peptides and proteins and cannot be ignored in designing NPs for various intracellular applications (Sée et al. 2009). The organelle distribution of NPs can be influenced by the uptake mechanism as well (Taylor et al. 2010). For example, ligand-free GNPs (positively charged) are internalized through a nonendosomal uptake mechanism and show localization of NPs in cytosol instead of endosomes and lysosomes (Taylor et al. 2010). Coincubation at 4°C didn't inhibit NP-uptake, suggesting diffusion as a possible entrance mechanism instead of endocytosis. PLGA NPs seem to behave differently than the other NP systems discussed so far. Once internalized through endocytosis process, PLGA NPs are distributed in endosomes, lysosomes, cytosol, ER, and Golgi complex (Panyam et al. 2002, Cartiera et al. 2009). For PLGA NPs, the mechanism of rapid escape into cytosol is believed to be selective reversal of the surface charge of NPs in the endolysosomal compartment, which causes the NPs to interact with the endo-lysosomal membrane and escape into the cytosol (Panyam et al. 2002). Based on these studies, it is possible to modify the surface of the NPs to tailor the intracellular processing of NPs for efficient delivery of variety of therapeutic agents, such as DNA and drugs. However, more studies are required to fully understand intracellular processing mechanisms of NPs based on the uptake mechanism, surface properties, size, shape, and NP-material.

Exocytosis process of NPs

Both the endocytosis and exocytosis of NPs are energy-dependent processes (Panyam et al. 2002, Chithrani & Chan 2007). Using TEM images, the endocytosis and exocytosis processes of GNPs are illustrated in Figure 1B. Several research groups studied exocytosis of NPs quantitatively and qualitatively. The NP uptake increased with incubation time in the presence of NPs in the medium; however, once the extracellular NP concentration gradient was removed, exocytosis of NPs occurred with about 65% of the internalized fraction undergoing exocytosis in 30 to 40 minutes (Panyam et al. 2002, Chithrani & Chan 2007, Wilhelm et al. 2008). Smaller GNPs appeared to exocytose at a faster rate and at a higher percentage than larger NPs. However, there is a remarkable difference between the percentages of exocytosis of GNRs versus spherical GNPs. GNRs are exocytosed faster than GNPs. Fast processing and excretion of GNRs can be attributed to the fewer number of proteins and receptors on their surfaces. Fast exocytosis process has also been reported for carbon nanotubes as well (Jin et al. 2008).

This review is so far mainly focused on the intracellular fate of NPs internalized through endocytosis process. However, Brust and co-workers have demonstrated that the NPs can be surface modified to cross the cell membrane avoiding endocytosis process for improving nuclear delivery (Nativo et al. 2008). The nuclear targeting of NPs in live cells is generating widespread interest because of the prospect of developing novel diagnostic and therapeutic strategies such as gene therapy (Tkachenko et al. 2003, Tkachenko et al. 2004, Suh et al. 2005, Oyelere et al. 2007, Nativo et al. 2008). However, the intracellular fate of NPs intended specially for nuclear delivery is still not well understood. The next section will give a brief overview of recent research work related to nuclear targeting of NPs.

Intracellular fate of gold nanostructures targeted for nuclear delivery

NPs designed for targeted nuclear delivery must be able to bypass or escape the endo-lyso pathway. One common approach to targeted nuclear delivery is the conjugation of drug molecules and NPs to nuclear targeting peptides for investigation of their transport (Feldherr et al. 1992, Tkachenko et al. 2004, De La Fuente & Berry 2005, Oyelere et al. 2007). GNPs are being used in this regard due to their small size, ease of preparation, strong absorbing and scattering properties, as well as their biocompatibility (Turkevich et al. 1951, Tkachenko et al. 2003, Connor et al. 2005, Oyelere et al. 2007, Rayavarapu et al. 2007). Tkachenko et al. have conjugated synthetic cellular targeting peptides to nanometer-sized GNPs through bovine serum albumin (BSA) protein and demonstrated the nuclear delivery of NPs by using video-enhanced colour differential interference contrast microscopy (Figure 6A). El-sayed and coworkers have conjugated peptides directly onto GNRs for nuclear targeting (Oyelere et al. 2007). Most of the NPs discussed so far were internalized through an endocytosis process. Brust and co-workers have shown that the endosomal pathway of these peptide- GNP complexes can be avoided significantly by appropriate modification of the particles with so-called cell penetrating peptides (CPPs) to cross the barriers of intact cells (Nativo et al. 2008). TEM images of cells with GNPs localized in the nucleus as well as in the cytoplasm are shown in Figure 6B. The size of the NPs plays an important role in nuclear targeting as well and the diameter of the NPs has to be less than 30 nm for import through the nuclear pore complex (Berry et al. 2007). However, more studies are needed to fully understand the mechanisms of cell uptake and intracellular fate of NPs conjugated for nuclear delivery.

Figure 6. Nuclear targeting of GNPs and GNRs. (A) Peptide-modified GNP complexes. Left: Illustration of the synthesis of peptide-modified colloidal GNP conjugates; Right: Visualization of targeted NPs in HepG2 cells by video-enhanced color differential interference microscopy. (B) PEG-modified GNPs functionalized with a combination of CPPs and NLS. Left: TEM image showing NPs localized in both nucleus and cytosol; Right: Variation of NPs in nucleus for a cell population. Reproduced with permission from Tkachenko et al. 2004 and Nativo et al. 2008.

Applications of gold nanostructures

Gold nanostructures can be used as a model system to understand how physiochemical properties affect the intracellular fate of NPs as summarized in this review. However, Gold nanostructures have now being used extensively for imaging, drug delivery, and therapeutics in cancer therapy (Supplementary Section S4 – online only). The potential use of Gold nanostructures as a contrast reagent for imaging and selective photothermal therapy of cancer cells has been demonstrated by several research groups (El-Sayed et al. 2005, Loo et al. 2005, El-Sayed et al. 2006, Huang et al. 2006). SERS and PAT (Photoacoustic Tomograpy) imaging have recently been used for detecting GNPs under in vivo conditions (El-Sayed et al. 2005, Qian et al. 2008, Zhang et al. 2009a). It has been shown that systemically administered cancer drug conjugated GNPs could improve the survival rates in tumour-bearing mice (Paciotti et al. 2010). Combination therapy, or the use of multiple drugs, has been proven to be effective for complex diseases, but the differences in chemical properties and pharmacokinetics can be challenging in terms of loading, delivering, and releasing multiple drugs. As a first step towards overcoming this challenge, Hamad-Schifferli and co-workers have demonstrated that it is possible to load and selectively release two different DNA oligonucleotides from two different GNRs by matching laser excitation wavelength to the nanorods' infrared SPR (Figure 7A) (Wijaya et al. 2009). It was also shown that GNRs can be used to generate heat by laser irradiation and remotely release the encapsulated materials at the site of interest (Skirtach et al. 2005, Skirtach et al. 2006, Volodkin et al. 2009). In radiation medicine, GNPs have been used as an imaging contrast agent and also to enhance DNA damage induced by anti-cancer drugs and radiation (Hainfeld et al. 2004, Hainfeld et al. 2006, Zheng & Sanche 2009, Brown et al. 2010). A recent review article summarizes the recent developments in use of Gold nanostructures in cancer nanotechnology (Cai et al. 2008). Gold nanostructures will play a bigger role not only as a model system to study physiochemical dependent biological fate of NPs but also as a therapeutic agent, as an imaging contrast agent, and as a drug delivery vehicle in the future.

Figure 7. Applications of GNPs. (A) Schematic explaining selective release of drugs. DNA-conjugated GNRs with different aspect ratios for selective release of drug molecules using laser irradiation. (B) GNPs incorporated multifunctional NP-system with both therapeutic and imaging capability. Reproduced with permission from Wijaya et al. 2009.

Conclusion

Detailed studies on the interface of nanostructures with biological systems provide guidance for proper design of multifunctional NPs with both imaging and therapeutic capability. In this review, Gold nanostructures were used as a model system to discuss how the size, shape, and surface properties of NPs affect their uptake and intracellular fate in vitro. However, further studies are needed to fully understand how physiochemical properties of NPs affect their biological fate both in vitro and in vivo for proper designing of multifunctional NPs for applications. In order to built NP-based systems with both therapeutic and imaging capabilities, GNPs are already being incorporated into other polymer or lipid-based NP systems (Skirtach et al. 2006, Sung-Hee et al. 2006, Chithrani et al. 2009a). In addition, the core and surface of polymer or lipid-based NPs can be used to contain imaging probes, anti-cancer drugs, and target ligands to facilitate combinational therapy and improve imaging capability (Figure 7B). GNPs have also been shown to be non-toxic to human cells (US National Cancer Institute-06-C-0167) (Connor et al. 2005). Hence, the clinical perspectives of multifunctional NPs are promising. However, gold nanostructure-based platforms are still at the initial stage of development and much more research is required before they can be applied in clinical applications.

Acknowledgements

I would like to thank Prof. David A Jaffray, Prof. Richard P Hill, Dr. Geof Aers, and Mr. Nicolas Gonzalez for their assistance in preparing this article. I would also like to acknowledge the following institutions: Canadian Institute for Health Research (CIHR), Ontario Institute for Cancer Research (OICR), Fidani Radiation Physics Centre at Princess Margaret Hospital, University Health Network, and STTARR innovation centre at Medical Discovery Tower, Toronto, Canada.

Declaration of interest: The author reports no conflict of interest. The author alone is responsible for the content and writing of the paper.