Multifunctionality of lipid-core micelles for drug delivery and tumour targeting

Abstract

Phospholipid micelles have proven to be the versatile pharmaceutical nanocarrier of choice for the delivery of poorly soluble chemotherapeutics for cancer therapy using various treatment modalities. Phospholipid micelles are typically expected to increase the accumulation of the loaded drugs in tumour tissues by taking advantage of the enhanced permeability and retention effect and by ligand-mediated active targeting. Furthermore, by tailoring the composition of the micelles, it is possible to enhance the intracellular delivery of the cargo. This review highlights the important advancements in our laboratory with polyethyleneglycol phosphatidylethanolamine (PEG-PE)-based micellar drug delivery systems for improvement of the therapeutic efficacy of poorly soluble anticancer drugs.

Keywords::

• Phospholipid micelles
• cancer targeting
• stimuli-sensitive nanocarriers

Introduction

The delivery of therapeutically active molecules to a desired target site remains a challenging task. Poor aqueous solubility of a drug and its unwanted distribution to healthy tissues are often responsible for the limited effectiveness of treatments (van Zuylen et al. 2001). Numerous nanocarrier-based drug delivery and drug-targeting systems have been devised in attempts to minimize drug degradation and inactivation upon administration, prevent undesirable side-effects, and increase drug bioavailability and the fraction of drug delivered to the target (Peer et al. 2007, Mozafari et al. 2009).

An important mechanism for the increased accumulation of many macromolecules and nanocarriers in various pathological areas such as tumours and infarcts was explained as a result of a so-called enhanced permeability and retention (EPR) effect (Maeda et al. 2000, 2001, Maeda 2001). This effect is a result of leaky blood vessels that allow an increased penetration of large molecules and even for small particles which then extravasate and accumulate in an interstitial space. Such an accumulation is often facilitated by the impaired lymphatic drainage (Maeda et al. 2001).

Among various pharmaceutical nanocarriers, liposomes, micelles, and polymeric nanoparticles have been the most extensively studied. Compared to low molecular weight free drug, drugs loaded in these nanocarriers usually exhibit favourable pharmacokinetics, biodistribution, and toxicity profiles (Liu et al. 2006, Bae and Kataoka 2009). Each of these nanocarrier types has characteristic features such as adequate drug loading, a suitability for surface-modification with various functionalities such as monoclonal antibodies and other moieties such as folate, transferrin, etc. (Torchilin 2001, 2005b, 2007). In certain special cases these nanocarriers can be made stimulus-sensitive and thus respond to local stimuli characteristic of the pathological site. For example, releasing an entrapped drug upon a change in local pH or temperature provided by surface-attached pH- or by temperature sensitive components (Bae and Kataoka 2009, Torchilin 2009).

Polymeric micelles are a special type of supramolecular assembly formed by self-assembling amphiphilic block-copolymers that consist of hydrophilic and hydrophobic monomer units in aqueous media and especially useful for serving as delivery system for hydrophobic poorly soluble drugs (Kataoka et al. 2001, Torchilin 2001, Mikhail and Allen 2009). They undergo dynamic physico-chemical changes during drug entrapment and release by molecular assembly and dissociation between the block copolymer components. Polymeric micelles exhibit a typical core-shell structure with sizes of 10–100 nm. Micelle size usually depends both on the chemical structure of the drugs and type of micelle core. At low concentration in an aqueous medium, these components exist as separate units. However, as their concentration is increased, aggregation takes place. The concentration of a monomeric amphiphile at which micelles form is called the critical micelle concentration (CMC). Following i.v. administration and dilution of micelles, a significant challenge can result in terms of achievement of adequate thermodynamic stability of the micellar system in vivo. However, kinetic analysis of this micelle-to-unimer exchange demonstrates that it is much slower compared to the low molecular weight surfactant micelles and depends on the several properties of the core-forming block including its glass transition temperature and melting temperature (Kang et al. 2005).

The CMC varies depending on block copolymer type and composition. It is generally in the range of 10^-6 and 10^-7 M, which is 1000 times lower than that of low molecular weight surfactants (Kwon et al. 1994). The major driving force behind self-association of amphiphilic polymers is the decrease in free energy of the system due to removal of hydrophobic fragments from the aqueous surroundings resulting from the formation of a hydrophobic micelle core stabilized with hydrophilic blocks exposed to water (Martin 1993, Kwon and Okano 1999). In certain cases, some additional interactions such as metal complexation (Nishiyama and Kataoka 2001) and electrostatic interaction (Harada and Kataoka 1995) may share responsibility.

When used as drug carriers in aqueous media, micelles will solubilize molecules of poorly soluble non-polar pharmaceuticals within the micelle core. Polar molecules can be adsorbed on the micelle surface, and substances with intermediate polarity distribute along the surfactant molecules in intermediate positions (Torchilin 2001). The micellar solubilization of a drug increases water solubility of a sparingly soluble drug (Lin and Kawashima 1985, 1987, Alakhov and Kabanov 1998), improves its bioavailability (Alakhov and Kabanov 1998), reduces toxicity and other adverse effects (Matsumura et al. 1999), enhances permeability across physiological barriers, and produce useful changes in biodistribution (Yokoyama et al. 1990, Kwon et al. 1993). Ideally, the core compartment of a pharmaceutical polymeric micelle should have a high loading capacity, a controlled release profile of the incorporated drug, and good compatibility between the core-forming block and incorporated drug. The micelle corona provides effective steric protection for the micelle core and determines the micelle hydrophilicity, charge, depending on the length and surface density of its hydrophilic blocks, and the presence of reactive groups suitable for further micelle derivatization, such as attachment of targeting moieties (Torchilin, 2001).

Beyond solubilizing hydrophobic drugs, block copolymer micelles can target their payload to specific tissues both passively and actively. Passive targeting is due primarily to the small micellar size which permits spontaneous penetration into the interstitium of tissue compartments with a relatively leaky vasculatures (particularly tumours and infarcts) by the EPR effect (Maeda et al. 2000, 2001, Maeda 2001). Active targeting can be achieved by surface attachment of target-specific molecules (Torchilin 2001, 2007). Figure 1 shows a typical scheme for micelle formation in aqueous medium from an amphiphilic molecule, its loading with a poorly soluble drug, and provides examples of further ways to modify the micelle to improve its performance as a pharmaceutical carrier.

Figure 1. Pharmaceutical micelles. Spontaneous micelle formation from amphiphilic molecules in aqueous media and loading with hydrophobic drug (A), a multifunctional pharmaceutical micelle containing; (a) specific targeting ligand, usually a monoclonal antibody such as 2C5, attached to the micelle surface [Gao et al. 2003]; (b) heavy metal atom such as ¹¹¹In, or Gd loaded onto the micelle via the micelle-incorporated chelating moiety for gamma- or MR imaging application [Torchilin 2001]; (c) cell-penetrating peptide, CPP, such as TATp attached to the micelle surface and allowing for the enhanced uptake by the cells [Sawant and Torchilin 2009] (B).

Characteristic properties of polymeric micelles such as drug loading ability and in vivo fate are influenced by the design of the amphiphilic block copolymers. The ideal polymeric micelles should be able to load, protect, deliver and release the entrapped cargo to the targeted site with a desirable pharmacokinetic profile. Within the structure of an amphiphilic polymer, monomer units with different hydrophobicities can be combined randomly, and be represented by two conjugated blocks consisting of the same type (A-B type copolymers), or be made from alternating blocks with different hydrophobicities (A-B-A type copolymers). Alternatively, the hydrophilic backbone chain of a polymer can be grafted with hydrophobic blocks (graft co-polymers) (Jones and Leroux 1999, Torchilin 2001). The hydrophilic shell is responsible for micelle stabilization and the interaction with plasma proteins and cell membranes. The shell usually consists of poly(ethylene glycol) (PEG) blocks with a molecular weight of 1–15 kDa. Other polymers such as poly(N-isopropylacrylamide) (Chung et al. 1999) and poly(alkylacrylic acid) (Chen et al. 1995) impart temperature or pH-sensitivity to the micelles. The hydrophobic core generally consists of a biodegradable polymer such as poly(β-benzyl-L-aspartate) (Kataoka et al. 2000, Lavasanifar et al. 2002, Kawano et al. 2006, Watanabe et al. 2006), poly(DL-lactide) (Pierri and Avgoustakis 2005) or poly (ϵ-caprolactone) (Forrest et al. 2006, Savic et al. 2006), or non-biodegradable polymer such as polystyrene or poly(methyl methacrylate) (Inoue et al. 1998). The core may also consist of a water soluble polymer (e.g., poly(aspartic acid)) which is rendered hydrophobic by the chemical conjugation of a hydrophobic drug (Yokoyama et al. 1992). In some cases, phospholipid residues can be used as hydrophobic core-forming compounds (Torchilin 2001, Gao et al. 2002, Lukyanov and Torchilin 2004). Polylactone-PEG double and triple block copolymers (Lin et al. 2003) have been suggested for use as micelle-forming polymers as well as poly(2-ethyl-2-oxazolineblock-poly(epsilon-caprolactone), which forms 20 nm micelles with good loading of paclitaxel (Cheon Lee et al. 2003). Chitosan-grafted with hydrophobic groups such as palmitoyl, has become popular for pharmaceutical micelle preparation due to its high biocompatibility (Jiang et al. 2006). Dendrimeric micelles, such as biaryl-based ones, have also been suggested (Ambade et al. 2005). New materials for pharmaceutical micelles include both, new copolymers of PEG (Prompruk et al. 2005) and completely new macromolecules, such as scorpion-like polymers (Djordjevic et al. 2005) and other star-like and core-shell constructs (Arimura et al. 2005). Micellar compositions of various drugs have been suggested for parenteral (Le Garrec et al. 2004, Shuai et al. 2004, Soga et al. 2005), oral (Park et al. 2005, Mathot et al. 2006), nasal (Gao et al. 2006) and ocular (Pillion et al. 1996, Lallemand et al. 2003) application.

Polymeric micelles can be prepared by two methods. The choice of method is usually determined by the solubility of the micelle-forming block copolymer in aqueous media. In the dialysis method, the block copolymers are dissolved in organic solvents such as dimethylsulfoxide, dimethylformamide, acetonitrile, followed by dialysis in aqueous medium. In a second method, block copolymers are dissolved in organic solvents such as cholorform, methanol, dichloromethane followed by removal of solvents by rotary evaporation to form a thin, dry film of block copolymers in a round bottom flask. The micelles are formed upon hydration of this film with aqueous solutions. Poorly soluble drugs can be physically entrapped during the process of micelle formation or chemically conjugated to block copolymer units. Examples of drugs loaded into various polymeric micelles are presented in Table I.

Table I. Examples of various drug-loaded polymeric micelles.

Block copolymer	Drugs incorporated
Pluronics	Doxorubicin, cisplatin (Alakhov et al. 1996); epirubicin (Batrakova et al. 1996); haloperidol (Kabanov et al. 1989); ATP (Slepnev et al. 1992)
Polycaprolactone-b-PEG	FK506, L-685,818 (Allen et al. 1998)
Polycaprolactone-b-methoxy-PEG	Indomethacin (Shin et al. 1998)
Poly(N-isopropylacrylamide)-b-PEG	Miscellaneous (Chung et al. 1998)
Poly(γ-benzyl-L-glutamate)-b-PEG	Clonazepam (Jeong et al. 1998)
Poly(β-benzyl-L-aspartate)-b-PEG	Doxorubicin (Kwon et al. 1993, Kataoka et al. 2000); indomethacin (La et al. 1996); KRN (Yokoyama et al. 1998a, Yokoyama et al. 1998b); amphotericin B (Yu et al. 1998)
Poly(L-lysine)-b-PEG	DNA (Katayose and Kataoka, 1997; Katayose and Kataoka, 1998)
PEG-PE	Dequalinium (Weissig et al. 1998a); soya bean trypsin inhibitor (Weissig et al. 1998b); camptothecin (Li et al. 2005); paclitaxel (Gao et al. 2002); porphyrine (Gao et al. 2002); (Wang et al. 2004)
Poly(D,L-lactide)-b-methoxy-PEG	Paclitaxel (Uchino et al. 2005); testosterone, griseofulvin (Kim et al. 2004)
Poly(N-vinyl-2-pyrrolidone)-b-poly(D,L-lactide)	Indomethacin (Benahmed et al. 2001)
Poly(N-isopropylacrylamide)-based micelles (pH-sensitive)	Phtalocyanine (Dufresne et al. 2004)
Dextran-b-poly(epsilon-caprolactone) diblock copolymer (Dex-SS-PCL) (reduction sensitive)	Doxorubicin (Sun et al. 2010)
Poly(L-histidine)-b-PEG (folate-targeted)	Doxorubicin (Lee et al. 2003)
PEG-b-poly(lactic-co-glycolic acid) (cyclic RGD peptide-modified)	Paclitaxel (Zhan et al. 2010)
mPEG-block-poly(delta-valerolactone) (epidermal growth factor receptor targeted)	CM-DiI (fluorescent probe) (Zeng et al. 2006)
Chitosan grafted with palmitoyl	Ibuprofen (Jiang et al. 2006); puerarin (Weiping et al. 2006)

Polymeric micelles that we are currently most interested in are the lipid-core-micelles. The present review summarizes the work carried out in our laboratory on polyethylene glycol-phosphatidylethanolamine (PEG-PE) micelles.

Lipid-core PEG-PE-based micelles as an example

The use of lipid moieties as hydrophobic blocks capping hydrophilic polymers such as PEG chains provides additional advantages for particle stability when compared with conventional amphiphilic polymer micelles. The presence of two fatty acid acyls can contribute considerably to an increase in the hydrophobic interactions between the polymeric chains in the micelle core (Lukyanov and Torchilin 2004). The chemical structure of PEG-PE is shown in Figure 2.

Figure 2. Chemical structure of PEG-PE, n = 14–18.

All versions of PEG-PE conjugates form micelles with a spherical shape, uniform size distribution (7–35 nm) and very low CMC values (in a high nanomolar to low micromolar range) because of the strong hydrophobic interactions between the double acyl chains of the phospholipid residues (Table II).

Table II. Particle diameter and CMC values of PEG-PE micelles.

Micelle-forming conjugate	Particle size (nm)	CMC (M)
PEG750-DSPE	7–15	1×10^−5
PEG2000-DSPE	7–20	1×10^−5
PEG5000-DSPE	10–40	6×10^−6
PEG2000-DOPE	7–20	9×10^−6
PEG5000-DOPE	10–35	7×10^−6

It is important to note that after the entrapment into micelles, the pharmacokinetic profile of the entrapped drug is dependent on the extent to which the drug remains encapsulated in the micelle following administration. Ideally the polymeric micellar system should function as a solubilizer and demonstrate good stability and drug retention in vivo in the presence of blood components. Micelles with very poor stability following i.v. injection will not alter the pharmacokinetic profile of the drug itself and thus function mainly just as solubilizers (Kim et al. 2001). In this regard, it has been observed that PEG₂₀₀₀-PE and PEG₅₀₀₀-PE micelles retain their size characteristics after 48 h incubation in the blood plasma (Lukyanov et al. 2002). This suggests the stability of the PEG-PE micelles in biological fluids upon parenteral administration.

PEG-PE micelles are structured in such a way that the outer PEG corona, known to be highly water soluble and highly hydrated, serves as an efficient steric protector in biological media. On the other hand, the phospholipid residues, which represent the micelle core are extremely hydrophobic and can solubilize various poorly soluble drugs including paclitaxel (Gao et al. 2002), camptothecin (Li et al. 2005), porphyrine (Roby et al. 2006) and vitamin K3 (Wang et al. 2004). Micelles prepared from PEG-PE conjugates with shorter versions of PEG are more efficient carriers of poorly soluble drugs because of their greater hydrophobic-to hydrophilic phase ratio and can be loaded with drug more efficiently on a weight-to-weight basis (Torchilin 2001).

Retention of hydrophobic drugs within the core of stable PEG-PE micelles is critical to the development of true site-specific delivery vehicles. To demonstrate that the incorporated drug is firmly associated with micelles, PEG-PE micelles loaded with several drugs were dialyzed against an aqueous buffer in a volume sufficient to provide sink conditions. All tested preparations retained more than 90% of encapsulated drug within first 7 h of incubation. The micelles retained 95%, 75% and 87% of initially incorporated chlorine e6 trimethyl ester, tamoxifen and paclitaxel, respectively, even after 48 h incubation (Gao et al. 2002). These results confirm the high stability of PEG-PE micelles themselves and demonstrate the strong drug association with these micelles.

Various attempts have been made to further increase the solubilization efficiency of micelles by forming mixed micelles with the addition of another surfactant or hydrophobic material. One should expect that these mixed micelles will allow better solubilization of poorly soluble drugs because of some loosening of the micelle core (e.g., with egg phosphatidyl choline) (Gao et al. 2003) or an increased volume of the lipophilic core of mixed micelles (e.g., with D-α-tocopheryl polyethylene glycol 1000 succinate/vitamin E/Pluronic) (Arimura et al. 2005, Dabholkar et al. 2006, Li and Tan 2008, Sawant et al. 2008b).

PEG-PE micelles and passive targeting

PEG-PE micelles have been shown to have an increased accumulation in solid tumours due to the EPR effect (Lukyanov et al. 2002) (Figure 3). To achieve a significant EPR effect, an extended circulation time is required. Therefore, passive targeting relies on both, the size of the delivery system and the physico-chemical properties (mainly surface properties) that influence the circulation time of the micelles. The prolonged circulation provides micelles a better chance to accumulate in a target (Gabizon 1995). The circulation half-lives of i.v. PEG-PE micelles were from 1.2 to 2.0 h in mice (Lukyanov et al. 2002). Thus, the PEG-PE micelles clearly demonstrate longevity. The peak tumour accumulation time of PEG-PE micelles prepared from all versions of PEG-PE conjugates observed was at about 5 h post-injection. The increase in the size of PEG blocks increased the micelle circulation time. This could be due to better steric protection against opsonin penetration to micelle core. However, micelles prepared from PEG-PE conjugates with shorter versions of PEG, might be more efficient carriers of poorly soluble drugs because of their greater hydrophobic-to-hydrophilic phase ratio and their more efficient drug loading on a weight-to-weight basis.

Figure 3. Schematics of possible ways for PEG-PE-based micellar nanocarrier-mediated cancer therapy. Following the accumulation in a tumour due to EPR effect, micelles can be taken up by cells via endocytosis and release drug. In case of photodynamic therapy (PDT) micellar photosensitizer exerts its cytotoxic effect due to generation of reactive oxygen species (ROS) upon activation at a specific region (tumour) with a laser. Micellar nanocarriers can be made ‘targeted’ by attaching tumour cell-specific ligands to their surface.

Since transport across the tumour vasculature depends somewhat on the particular type of tumour (Hobbs et al. 1998), the use of micelles as drug carriers can be particularly useful for tumours, whose vasculature has a low cutoff size (below 200 nm). Thus, 15–20 nm PEG-PE micelles effectively delivered a model protein drug to a solid tumour with a very low cut off size (Lewis lung carcinoma) in mice (Torchilin et al. 2001b), while even small, 100 nm, long-circulating liposomes did not show an increased accumulation of their liposome encapsulated drug (Parr et al. 1997). Other recent data clearly indicates enhanced uptake of PEG-PE-based micelles by other experimental tumours (Torchilin et al. 2001b) in mice as well as into ischemic areas of the heart in rabbits with an experimental myocardial infarction (Lukyanov et al. 2004).

PEG-PE micelles and ligand-mediated targeting

Polymeric micelles can be surface-modified with moieties with an affinity for cellular receptors or components that are present on and/or upregulated by tumour cells (Figure 3). Various ligands such as sugar moieties (Jeong et al. 2005), epidermal growth factor (Zeng et al. 2006, Fonge et al. 2010), antibodies (Torchilin et al. 2003, Sawant et al. 2008b, Skidan et al. 2008), folate (Kim et al. 2009) and peptides (Nasongkla et al. 2004) have been tested for active targeting of polymeric micelles to tumours.

We have successfully prepared and analyzed tumour-targeted PEG-PE micelles by attachment of the anti-cancer nucleosome-specific monoclonal antibody 2C5 (mAb 2C5) to their outer shell, which recognizes the surface of a broad variety of tumour cells via tumour cell surface-bound nucleosomes (Gao et al. 2003, Roby et al. 2006, Sawant et al. 2008b, Skidan et al. 2008). These immunomicelles were prepared using PEG-PE conjugates with the free PEG terminus activated with a p-nitrophenylcarbonyl (pNP) group (Torchilin et al. 2001a). Diacyllipid fragments of such a bifunctional PEG derivative firmly incorporate into the micelle core, while the water-exposed pNP group, stable at pH values below 6, interacts with amino-groups of various ligands (antibodies and their fragments or peptides) at pH values above 7.5 to yield a stable urethane (carbamate) bond. To prepare immunomicelles, the corresponding antibody is incubated with the drug-loaded pNP-PEG-PE-containing micelles at a pH around 8. ELISA studies confirmed the retention of antibody activity after surface modification of the micelles. Approximately 10–20 antibody molecules were attached to a single micelle (Gao et al. 2003, Torchilin et al. 2003). The antibody modification does not increase the micelle size. Both micelles with and without antibody modification were spherically shaped and had a uniform size of about 20 nm (Torchilin et al. 2003).

In vitro cell binding studies demonstrated that rhodamine-labeled 2C5-immunomicelles bind effectively to the surface of various cancer cells including LLC (Lewis lung carcinoma), EL4 (T lymphoma), BT20 and MCF-7 (breast adenocarcinoma) cells compared to antibody-free micelles (Figure 4) (Gao et al. 2003).

Figure 4. Fluorescence microscopy of the binding of rhodamine-PE-labeled paclitaxel-loaded PEG-PE-based micelles to murine LLC and EL4 cells and to human BT-20 and MCF-7 cells. Cells were grown on cover slips to a confluence of 60–70%, incubated with various preparations for 1 h, washed and mounted cell-side down on glass slides using a fluorescence-free glycerol-based Trevigen^® mounting medium. For each pair, left images – phase contrast; right images – fluorescence. Modified from Torchilin (2005a).

Loading with an anticancer drug (such as paclitaxel or camptothecin) dramatically improved in vitro cancer cell killing by 2C5-modified micelles compared to antibody free micelles or free drug (Gao et al. 2003, Li et al. 2005, Sawant et al. 2008b). In vivo studies with LLC tumour-bearing mice showed an enhanced accumulation of paclitaxel-loaded 2C5-modified immunomicelles over plain micelles in the tumour both at 30 min and at 2 h post-injection. Thus, the mAb 2C5-immunomicelles brought higher amounts of paclitaxel to tumours compared to paclitaxel-loaded non-targeted micelles (Gao et al. 2003, Torchilin et al. 2003).

The 2C5-modified micelles have also been shown to have the promise for photodynamic therapy (PDT). PDT of cancer relies on the accumulation of a photosensitizing drug in the neoplastic tissue and subsequent formation of cytotoxic products by irradiation of the drug-containing tumour with a light of a suitable wavelength (Figure 3). In this regard, we have loaded the poorly soluble drug, meso-tetraphenylporphine (TPP) into PEG-PE micelles, additionally surface-modified with 2C5 antibody. These immunomicelles showed a stronger phototoxic effect against B-16 (murine melanoma) and MCF-7 cells in vitro when compared with TPP-loaded plain micelles at the same TPP concentration (Roby et al. 2006, Skidan et al. 2008). The cytotoxicity results were further confirmed by analysis of nuclear fragmentation resulting from the apoptosis associated with PTD-induced cell death in the presence of drug-loaded 2C5-immunomicelles observed with fluorescence microscopy of PDT-treated LLC cells stained with DAPI, a fluorescent dye (Figure 5). The cells treated with TPP-loaded mAb 2C5 immunomicelles clearly showed more DNA fragmentation (increased apoptosis) when compared to TPP-loaded plain micelles (control micelles). Apoptotic DNA fragments were faintly observed in the drug alone or light alone control groups (Figure 5). In vivo, the PDT treatment of subcutaneous melanoma-bearing C57BL/6 mice with 100 mW/cm² of 630 nm laser light at 9 h after the administration of the micellar TPP (1 mg/kg of TPP) resulted in a significant inhibition of tumour growth. Compared with controls, the weight of postmortem tumours was 3.5- and 7.5-fold smaller with TPP-loaded PEG-PE micelles and TPP-loaded PEG-PE 2C5-immunomicelles, respectively (Skidan et al. 2008).

Figure 5. Nuclear fragmentation following PDT of LLC cells with TPP-loaded PEG-PE micelles. An apoptotic cell is indicated by an arrow. Cells were stained with DAPI for visualization of nuclear fragmentation by fluorescence microscopy. Modified from Roby et al. (2006).

Intracellular delivery of micelles

Polymeric micelles cannot diffuse through the cell membrane but rather are internalized by endocytosis. Detailed reviews of the endocytotic pathways and endocytosis of nanocarriers can be found in (Mukherjee et al. 1997, Conner and Schmid 2003, Bareford and Swaan 2007). Following cell uptake, micelles are contained within acidic endosomes and are further directed to various transport pathways including fusion with lysosomes or exocytosis. Further improvement of the efficiency of drug-loaded micelles should involve enhancement of their intracellular delivery to compensate for excessive drug degradation in lysosomes as a result of the endocytosis-mediated capture of micelles by cells.

PEG-PE micelles carry a net negative charge (Lukyanov et al. 2004), which may hinder their internalization by cells. Modification of PEG-PE micelles with positively charged lipids may improve the uptake of drug-loaded micelles by cells. Such positively charged micelles could also more readily escape from endosomes and enter the cytoplasm of cells. To test these ideas, we have prepared paclitaxel-loaded micelles from mixture of PEG-PE and positively charged Lipofectin^® lipids (LL). The intracellular fate of paclitaxel-loaded PEG-PE/LL micelles and micelles prepared without the addition of the LL were investigated by fluorescence microscopy with BT-20 breast adenocarcinoma cells (Wang et al. 2005). Both fluorescently-labeled PEG-PE and PEG-PE/LL micelles were endocytosed by BT-20 cells (Figure 6). However, with PEG-PE/LL micelles, endosomes appeared to be partially disrupted and released drug-loaded micelles into the cell cytoplasm, result of the de-stabilizing effect of the LL component on the endosomal membranes. After 4 h incubation, larger, fused fluorescent endosomal structures became apparent in the case of LL-free micelles, whereas cells incubated with PEG-PE/LL micelles had smaller punctuate fluorescent structures in the cytoplasm. This could result in increased cytoplasmic delivery of paclitaxel. This delivery was supported by the results of in vitro cytotoxicity studies against BT-20 cells and A2780 cells (human ovarian carcinoma). The paclitaxel-loaded PEG-PE/LL micelles were significantly more cytotoxic compared to that of free paclitaxel or paclitaxel delivered using non-cationic LL-free PEG-PE micelles: in A2780 cancer cells, the IC50 values for free paclitaxel, paclitaxel in PEG-PE micelles, and paclitaxel in PEG-PE/LL micelles were 22.5, 5.8 and 1.2 μM, respectively. In BT-20 cancer cells, the IC50 values of the same preparations were 24.3, 9.5 and 6.4 μM, respectively.

Figure 6. Microscopy of BT-20 cells incubated with PEG-PE/ paclitaxel micelles and PEG-PE/LL/paclitaxel micelles for 2 and 4 h. Bright-field (left images in each pair) and fluorescence (right images in each pair). Arrows indicate fluorescent endosomes in cells incubated with PEG-PE/paclitaxel micelles for 2 h; partially degraded endosomes in cells incubated with PEG-PE/LL/paclitaxel micelles for 2 h; punctuate fluorescent structures in cells incubated with PEG-PE/LL/paclitaxel micelles for 4 h; larger (fused) endosomes in cells incubated with PEG-PE/paclitaxel micelles for 4 h. Modified from Torchilin (2005a).

The synthetic vitamin K3 (2-methyl-1,4-naphthoquinone, VK3) inhibits the growth of various cancer cell types both in vitro and in vivo (Nutter et al. 1991, Carr et al. 2002). Earlier, 1,8-diazabicyclo[5,4,0]undec-7-ene (DBU) was used to accelerate the formation of thioether analogs of VK3 that possess higher biological activity (Nishikawa et al. 1995). It was hypothesized that the arylation of cellular thiols by the VK3 and the acceleration of thioether formation by DBU would result in synergistic anticancer effects, when VK3 and DBU are delivered simultaneously inside cancer cells. With this in mind, we prepared PEG-PE polymeric micelles loaded with both VK3 and DBU (M-VK3-DBU) or only with VK3 (M-VK3) and their action on various cancer cell lines in vitro was investigated. The endocytosis patterns of fluorescently labeled (rhodamine-PE) M-VK3-DBU and MVK3 micellar preparations are shown in Figure 7. Both micellar preparations attached to the surface of BT-20 cells within the first 30 min of incubation. At 1 h, both M-VK3-DBU and M-VK3 were endocytosed and could be seen in the endosomes of the cell (fluorescent punctuate signals inside cells denoted by arrows). Some broader structures with less compact fluorescence could also be seen in the cytoplasm of BT-20 cells incubated with M-VK3-DBU but not with M-VK3. At 4 h, an accumulation of the fluorescence in the perinuclear space was observed (fluorescence denoted by arrows) within the cells incubated with M-VK3-DBU. However, no similar structures or fluorescence accumulation were found on the nuclear membrane in the BT-20 cells incubated with MIC-VK3. These results suggest that DBU can modify certain protein endosomal components and induce destabilization of endosomes, which lead to the escape of drug-loaded micelles directly into the cytoplasm. This cytoplasmic escape was supported by the significantly enhanced cytotoxicity of M-VK3-DBU towards several cancer cell lines compared to that of M-VK3. At an equal VK3 concentration, ID50 values for BT-20, A2780, and murine Lewis lung carcinoma (LLC) tumour cells were 3.6, 10.5, and 5.7 μM (as VK3), respectively, for M-VK3, and 1.9, 4.2, and 3.3 μM, respectively, for M-VK3-DBU.

Figure 7. Fluorescence microscopy of the interaction of rhodamine-PE-labeled micelles with BT-20 cells. Arrows indicate binding of micelles to the cell surface (a); formation of endosomes (b); endosomal escape (c) and accumulation in the perinuclear space (d). Modified from Torchilin (2005a).

A promising approach for the intracellular delivery that has emerged over the last decade is the use of cell penetrating peptides (CPPs) (Schwarze et al. 1999). In particular, the CPP derived from the human immunodeficiency virus-1 transactivator protein, TAT peptide (TATp) has attracted much interest. TATp-mediated cytoplasmic uptake of polymers, plasmid DNA (Astriab-Fisher et al. 2002, Nguyen et al. 2008), nanoparticles (Lewin et al. 2000, Zhao et al. 2002, Rao et al. 2008), liposomes (Torchilin 2001, Torchilin et al. 2001b, Levchenko et al. 2003, Fretz et al. 2004) and micelles (Sethuraman and Bae 2007, Sawant et al. 2008b) has been reported. A variety of uptake mechanisms appear to be involved in different systems, and in some cases, the mechanism is cell-type or cargo-specific (Zorko and Langel 2005). Smaller molecules attached to TATp seem to transduce directly into cells by energy independent electrostatic interactions and hydrogen bonding (Vives et al. 2003), but the larger cargos get into cells by an energy dependent macropinocytosis pathway (Wadia et al. 2004). One of the major obstacles to use of TATp-assisted intracellular delivery of pharmaceutical nanocarriers is the lack of selectivity of TATp. This non-selectivity has generated concern about drug-induced toxic effects towards normal tissues. Thus, intratumoural administration of TATp containing nanocarriers may serve as good solution to this problem for delivery of anticancer drugs at least in certain cases. With this in mind, we prepared and studied paclitaxel-loaded TATp containing PEG-PE micelles. For an efficient interaction of the micelle-attached TATp with the cell, the TATp moiety should be located ‘above’ the carrier surface (Sawant et al. 2008a). Such, paclitaxel-loaded micelles were prepared using PEG₇₅₀-PE as the main micelle-forming component with the addition 2.5 mol % TATp-PEG₁₀₀₀-PE (Figure 8A).

Figure 8. Schematic representation of the TATp-modified micelles (A); In vitro interaction of rhodamine-PE labeled PEG-PE micelles with 4T1 cells. Left panel shows the bright field and right panel shows the fluorescent microscopy of 4T1 cells treated with rhodamine-labeled PEG₇₅₀-PE micelles (a), rhodamine-labeled PEG₇₅₀-PE micelles modified with TATp-PEG₁₀₀₀-PE (b). Magnification ×40 objective. Modified from Sawant and Torchilin (2009).

The in vitro cell interaction of the TATp-bearing PEG-PE micelles was confirmed by fluorescence microscopy with 4T1 cells (Figure 8B). Plain micelles composed of PEG₇₅₀-PE demonstrated limited interaction with the cells (Figure 8B, a). However, the use of the TATp-bearing PEG₇₅₀-PE micelles resulted in the expected strong interaction with the cells (Figure 8B, b). This enhanced interaction also resulted in increased in vitro cytotoxicity against MCF-7 and 4T1 cells with paclitaxel-loaded TATp-bearing micelles compared to paclitaxel-loaded micelles without TATp at both 5 and 50 nM paclitaxel concentrations.

In in vivo studies, to avoid any unwanted distribution of paclitaxel-loaded TATp-micelles, those were injected intratumourally in mice and tumours were harvested after 48 h. The nuclear DNA fragmentation in tumour sections undergoing apoptosis was observed using the TUNEL assay with a DNA fragmentation kit. The differentiation of apoptosis from necrosis occurs because DNA fragments produced by the apoptosis are exclusively labeled with the TUNEL method (Gold et al. 1994). The results of the TUNEL staining of tissue sections are shown in Figure 9. In DAPI stained images, it is difficult to observe differences among groups since the cell nuclei of live cells represented bright blue fluorescence attributed to DAPI staining. The cell nuclei of non-apoptotic bodies did not exhibit the green fluorescence attributed to FITC-labeled TdT. Very few TUNEL-positive cells were observed in tumours injected with free paclitaxel and paclitaxel-loaded micelles (Figure 9C). However, significant apoptotic cell death was observed in tumours treated with paclitaxel-loaded TATp-bearing micelles (Figure 9D).

Figure 9. Detection of apoptotic cells by fluorescence microscopy of frozen tumour sections. Apoptosis was determined by TUNEL. The left panel shows the sections stained with DAPI and the right panel shows TUNEL. Negative control (A), free paclitaxel (B), paclitaxel-loaded micelles without TATp (C), paclitaxel-loaded micelles with TATp (D). Magnification ×20 objective. Modified from Sawant and Torchilin (2009).

The major drawback of TATp-mediated intracellular delivery is the lack of selectivity of TATp, which has rendered this cellular drug delivery method unacceptable at present. This restriction has led to efforts to build ‘smart’ nanocarriers (Figure 10) in such a way that during the first phase of delivery the non-specific cell-penetrating function (in our case TATp) is shielded by functions that provide organ/tissue-specific delivery (with a sterically protecting polymer or targeting antibody). Upon the accumulation in the target, the protecting polymer or antibody attached to the surface of the nanocarrier via stimuli sensitive bond detaches in the presence local pathological conditions (abnormal pH or temperature) and exposes the previously hidden second function allowing subsequent delivery of the carrier and release cargo inside cells (Sawant et al. 2006).

Figure 10. Schematic of a ‘smart’ nanocarrier with a temporarily ‘hidden’ function, for example a CPP, and ‘shielding’ polymeric coat (with or without targeting antibody attached to it) providing longevity in the blood and specific target (tumour) accumulation and preventing the hidden function from premature interaction with target cells. Polymeric chains are attached to the carrier surface via low pH-degradable bonds. After the accumulation in the tumour due to PEG (longevity) and/or antibody (specific targeting), pH-dependent de-shielding of the temporarily hidden cell-penetrating function allow for carrier penetration inside tumour cells.

To test this idea, we prepared targeted PEG-PE-based micelles simultaneously possessing several functionalities (Sawant et al. 2006). First, the system was rendered capable of targeting a specific cell or organ by attaching a monoclonal antibody (infarct-specific anti-myosin antibody 2G4 or cancer-specific anti-nucleosome antibody 2C5) to the surface via reactive para-nitrophenyl (pNP)-PEG-PE moieties. Second, these micelles were additionally modified with TATp moieties attached to the surface of the nanocarriers by using TATp-(short PEG)-PE derivatives. PEG-PE used for micelle preparation was made degradable (detachable) by inserting a pH-sensitive hydrazone bond between PEG and PE (PEG-Hz-PE). At normal pH values, TATp functions on the surface of nanocarriers were ‘shielded’ by the long protecting PEG chains (pH-degradable PEG₂₀₀₀-PE or PEG₅₀₀₀-PE) or by the long pNP-PEG-PE moieties used to attach antibodies to the nanocarrier. At pH 7.5–8.0, micelles demonstrated high specific binding with antibody substrates, but very limited internalization by NIH/3T3. However, upon brief incubation (15–30 min) at low pH values (pH 5.0–6.0) these nanocarriers lost their protective PEG shell by acidic hydrolysis of PEG-Hz-PE and were effectively internalized by cells via exposed TATp moieties (Figure 11).

Figure 11. Fluorescence microscopy showing internalization of rhodamine-PE-labeled-TATp-containing micelles by NIH 3T3 fibroblast cells after preincubating micelles at pH 8.0 (A) and pH 5.0 (B)* for 30 min. *The pH of these formulations was raised back to pH 7.4 after their incubation at pH 5.0 and prior to incubation with cell. Modified from Sawant et al. (2006).

These ‘smart’ pH-sensitive micellar nanocarriers might have potential application for tumour-targeted delivery. When administered in vivo, the micelles will circulate in blood and accumulate in tumours via EPR effect and/or antibody-mediated targeting. The protective pH-sensitive PEG coating has the ability to sterically shield the TATp function and prevent it from the non-specific internalization while in circulation. It is known that intratumoural pH is slightly acidic (Vaupel et al. 1989). The TATp can become exposed inside the tumours due to removal of the PEG coat, which should result in the enhanced internalization of the micelles into target cells.

Summarizing, lipid-core polymeric micelles such as PEG-PE micelles represent a versatile drug delivery platform for broad variety of poorly soluble drugs. Each delivery system has its own challenges to face. Not much attention has been paid to cellular uptake and intratumoural distribution of PEG-PE micelles in vivo. Future studies, are necessary to understand the role of EPR effect and active targeting in micellar drug delivery.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.