Cell-Type Specific and Cytoplasmic Targeting of PEGylated Carbon Nanotube-Based Nanoassemblies

Abstract

In this paper we report the fabrication of a multivalent, cell-type specific and cytoplasmic delivery system based on single-walled carbon nanotubes. The latter were functionalized through adsorption of phospholipids terminated by biotinylated PEG chains functionalized with fluorochrome-coupled neutravidin, and subsequently with antibodies (anti-CD3ε and anti-CD28) for T cell receptor post-signaling endocytosis and a synthetic fusogenic polymer for disruption of lysosomal compartments. The biomimetic nanoassemblies were composed by PEGylated individual/very small bundles of carbon nanotubes having an average length and a standard deviation of 176 nm and 77 nm, respectively. The nanoassemblies were stably dispersed under physiological conditions, visible by conventional optical and confocal microscopy and specifically targeted to T cells both in vitro and in living animals. The addition of a fusogenic polymer to the nanoassemblies did not affect the cellular uptake and allowed the release into the cytosol of the targeted cells both in vitro and in the animals. The present manuscript is the first report about the cytoplasmic delivery of carbon nanotubes in a specific cell type in intact animals and paves the way for their use as in vivo intracellular delivery systems.

1. INTRODUCTION

An intracellular delivery system must achieve several objectives to successfully deliver a gene or drug cargo, namely, targeting to specific cell-surface receptors, internalization via the endosome/lysosome system, disruption of the endosomal/lysosomal vesicles to deliver the cargo to the cytoplasm or, eventually, to the nucleus. For systemic in vivo applications, such factors as physicochemical properties that prevent aggregation or nonspecific interactions with blood system components must also be considered. Interaction with blood components could form aggregates that obstruct capillaries and reduce the circulation time. Nanomedicine may lead to the development of more effective means for delivering and targeting pharmaceutical, therapeutic, and diagnostic agents.^1–3 The transport of several types of nanomaterials into adherent and nonadherent mammalian cell lines has been reported.^4–21 Nevertheless, many reported nanomaterials were characterized by surface chemistry that limited their capacity to interact multivalently with cell membrane receptors and the efficient targeting of specific cell types. Furthermore, independent of the mechanism (non-specific adsorptive endocytosis, phagocytosis, macropinocytosis, and receptor-mediated endocytosis), uptaken nanomaterials follow the same endo/lysosomal pathway, being transported from endosomes to lysosomes, the organelles that enzymatically digest macromolecules and are characterized by an acidic pH (approximately 4.5). The lysosomal vesicles subsequently fuse and assemble in the perinuclear region.²² Here the majority of nanomaterials remain without significant changes in distribution pattern. The endo/lysosome barrier to cytoplasmic entry represents a significant challenge in the use of nanomaterials as intracellular delivery systems.

Among recently nanotechnology-introduced materials, carbon nanotubes (NTs) have been intensely investigated because of such extraordinary physicochemical properties as high aspect ratios (length to diameter ratio), light weight, high mechanical strength, high electrical conductivity, high thermal conductivity, metallic or semi-metallic behavior, and high surface area.²³ These combined properties make NTs unique materials with the potential for nanomedical applications. Due to their extremely high aspect ratios, NTs represent a potential multivalent scaffold for drug or gene delivery and constructing nanoassemblies capable of interacting with cell membrane receptors. Two major obstacles to the use of pristine (non-functionalized) NTs as delivery vehicles are their poor solubility and tendency to aggregate under physiological conditions. To overcome these problems, NT surfaces have been functionalized by such techniques as adsorption, electrostatic interaction, or covalent bonding processes. Such functionalizations of NTs allowed their dispersion under physiological conditions,^10–15 conferred resistance to non-specific binding to biological species,²⁴ improved their (in vitro) cytotoxic profile,²⁵ and decreased their uptake by the liver or other organs of the reticuloendothelial system.^27–29 However, the chemistry currently used to functionalize NTs has limited their efficient targeting to specific cell types within a mixed population, the capacity to interact multivalently with cell membrane receptors and to reach a particular intracellular compartment. Only a few reports on the use NT-based nanoassemblies to target a specific cell line in vitro exist, whereas the use of such nanoassemblies to provide cytoplasmic transport of NTs into specific cell types in vivo has not been reported yet and could represent a fundamental step in paving the way for its use as a more effective means of treating malignancies.

NTs functionalized through adsorption of phospholipids terminated by poly(ethylene glycol) (PEG) chains have been used to deliver oligonucleotides into cultured cells^12–14 and to target integrin-positive tumors in mice.²⁹ In particular, PEGylated NTs (PNTs) exhibited long blood circulation times and low uptake by the reticuloendothelial system.²⁹ These results suggest that PNTs can be developed as far more effective nanomedicines than pristine NTs and other nanomaterials. The present work focuses on the generation of a multivalent, cell-specific and cytoplasmic delivery system constructed from PNTs functionalized with fluorescent neutravidin, and subsequently with antibodies for cell receptor post-signaling endocytosis, and a fusogenic polymer for disruption of lysosomal compartments (Scheme 1(A)). The biomimetic nanoassemblies were targeted to T lymphocytes both in vitro and in intact animals and delivered into the cytoplasm when engineered with a fusogenic polymer (Scheme 1(B)). Furthermore, the nanoassemblies were stably dispersed under physiological conditions, readily visible under light microscopy, biocompatible and functionalizable with intracellularly active proteins and nucleic acids by exploiting the strong affinity between biotin and free neutravidin on the PNT surface. Therefore, the reported PNT-based nanoassemblies could be used to achieve a specific cytoplasmic effect in targeted cells.

Scheme 1.

Cell-type specific and cytoplasmic targeting of PEGylated carbon nanotube-based nanoassemblies. (A) Functionalization of PEGylated carbon nanotubes with fluorophore-conjugated neutravidin, antibodies and a synthetic fusogenic polymer. Phospholipid, linked to biotinylated PEG chains (PL-PEGb), adsorbed onto carbon nanotubes (NTs) leaving their PEG chains protruding from the sidewalls (PNT). PNTs were hydrophilic, biocompatible and functionalized with Texas Red-conjugated neutravidin yielding nanoassemblies readily visible under light microscopy (PNT-TRNav). PNT-TRNav were functionalized with antibodies (anti-CD3ε and anti-CD28) for cell receptor post-signaling endocytosis and a synthetic fusogenic polymer [poly(2-propylacrylic acid)] for disruption of lysosomal compartments (PNT-TRNav-α₃α₂₈p). (B) T-lymphocytes membrane receptor post-signaling endocytosis and cytoplasmatic delivery in vivo. The nanoassemblies were injected intravenously into 129sv/J mice, delivered into T cells through membrane receptor post-signaling endocytosis, transported into secondary lysosomes through the fusion of endocytotic vescicles with primary lysosomes, and finally released into the cytoplasm through the disruption of the lysosomal compartments by the fusogenic polymer.

2. EXPERIMENTAL DETAILS

Preparation of Biotinylated Fusogenic Polymer.

Amino-terminated poly(2-propylacrylic acid), (Polymer Source, Inc., Dorval, Canada) (PPAA) (5 mg) and sulfosuccinimidyl-6-(biotin-amido)hexanoate (sulfo-NHS-LC-biotin, Pierce Biotechnology, Inc., Rockford, IL) (1 mg) were dissolved in DMF (50 μL) and PBS (2.7 mM KCl, 1.5 mM KH₂PO₄, 137 mM NaCl and 8.1 mM Na₂HPO₄, pH 7.4 from Mediatech, Inc., Herndon, VA) (4 mL), respectively. The two solutions were mixed and left to react at room temperature for 2 h. To maximize the yield of biotinylated polymer (PPAAb) another 1 mg of sulfo-NHS-LC-biotin was added and left to react for 2 h. PPAAb was purified from unlinked sulfo-NHS-LC-biotin by dialysis against PBS. The concentration of PPAAb (approximately 1 mg/mL) was determined using a FluoReporter biotin quantitation kit (Invitrogen Corp., Carlsbad, CA).

Preparation of Biotin-Functionalized PEGylated NTs (PNTs).

NTs (Carbon Solutions, Riverside, CA) (5 mg) in 10 mL of N-[biotinyl(polyethylene glycol)2000]1,2-distearoyl-sn-glycero-3-phosphoethanolamine (Avanti Polar Lipids, Alabaster, AL) (PL-PEGb, Scheme 1) (1 mg/mL) in PBS were ultrasonicated (Misonix, Inc., Farmingdale, NY) for 3 h. The mixture was centrifuged (Centrifuge 5417R, Eppendorf AG, Hamburg, Germany) at 32,000× g for 30 min to remove large NT bundles. The supernatant fraction was ultracentrifuged (Beckman Optima^™XL-80K Ultracentrifuge, Palo Alto, CA) at 100,000× g for 1 h. This second supernatant was further ultracentrifuged at 100,000× g for 2 h and the collected supernatant filtered 5 times through 100-kDa MWCO centrifugal filter devices (Ultrafree, Millipore Corp., Temecula, CA) to remove free phospholipids.

Functionalization of PNTs with fluorescent neutravidin (PNT-TRNav), antibodies (PNT-TRNav-α₃α₂₈) and a fusogenic polymer (PNT-TRNav-α₃α₂₈p). PNTs (50 μg) were incubated with TRNav (Texas Red-conjugated neutravidin, Pierce Biotechnology, Inc.) (2.5 mg) at room temperature for 2 h and then washed 3 times (centrifugation at 40,000× g for 30 min and redispersion of the pellet in PBS) in order to remove free proteins (PNT-TRNav). PNT-TRNav (40 μg) in 600 μL of PBS were incubated with 60 μg an equimolar mixture of biotinylated anti-CD3ε and anti-CD28 (eBioscience, Inc., San Diego, CA) for 1 h, then washed 3 times in PBS (PNT-TRNav-α₃α₂₈). Subsequently 50 μL of PPAAb were added and after 1 h the nanomaterials were washed 3 times in PBS (PNT-TRNav-α₃α₂₈p).

Characterization of PNT-Based Nanoassemblies.

UV-Vis absorbance (model 8453, Agilent Technologies, Palo Alto, CA, USA) and fluorescence (model MOS-250, Bio-Logic, Claix, France) spectroscopies were used for characterization. An atomic force microscope (5500 AFM, Agilent Technologies, Santa Clara, CA) was used to evaluate the length distribution of the PNT. A solution of PNT in PBS (10 μL) was dropped onto a freshly cleaved mica substrate (Ted Pella, Redding, CA). The droplet was allowed to stand for approximately 10 min at room temperature, and then the mica surface was rinsed with water and dried under a nitrogen stream. Images were taken in dynamic mode (acoustic alternate current, AAC) using Si₃N₄ tips (Nanosensors, Neuchatel, Switzerland).

In Vitro Cell Membrane Receptor-Mediated Internalization.

Jurkat T leukemia cells (or Raji lymphoma B cells) were grown at 37 °C in 5% CO₂ in media (RPMI-1640) supplemented with 10% fetal bovine serum (Tissue Culture Biologicals, Informagen, Inc., Newington, NH). Cells (10⁵) in logarithmic growth were washed with media, resuspended in 270 μL of media, incubated with PNT-based nanoassemblies (PNT-TRNav, PNT-TRNav-α₃, PNT-TRNav-α₃α₂₈, or PNT-TRNav-α₃α₂₈p) (250 ng) in PBS (30 μl) at 37 °C in 5% CO₂ atmosphere. For apoptosis evaluations incubated cells were washed twice with Annexin V buffer (10 mM HEPES/NaOH pH 7.4, 140 mM NaCl, 2.5 mM CaCl₂), resuspended in a solution of R-Phycoerythrin-conjugated Annexin V, incubated at 4 °C in dark for 20 min, washed and run on FACS machine, while, for necrosis evaluations, incubated cells were resuspended in a solution of propidium iodide and incubated at 4 °C for 10 min before running the sample on the cytometer.

In Vivo Cell Membrane Receptor-Mediated Internalization Cell Targeting.

Mice were injected intravenously with 200 μL of nanotube-based material and sacrificed after 24 h. For flow cytometry single cell suspensions were prepared from spleen by physical dissociaton between glass slides. Red blood cells were removed from spleen and blood samples by hypotonic lysis.

Flow Cytometry.

Prior to staining, non-specific Fc-receptor mediated binding of primary antibodies was blocked by incubating cells with anti-CD16/CD32 (eBioscience) according to the manufactures protocol. Cells were stained with anti-CD45R(B220) –PE (eBioscience) or –APC-Cy7 (BD Pharmingen), anti-CD3ε–FITC (eBioscience), and anti-CD11b–PE-Cy7 (eBioscience). Flow cytometric collection was performed using a BD FACSCanto 6 color flow cytometer with BD FACSDiva software. Stained cells and PNT⁺ cells were analyzed and displayed using FlowJo software (Tree Star, Inc). Total cell gates refer to single leukocytes determined by Forward Scatter (FSC) and Side Scatter (SSC) profile. Percent of PNT⁺ T cells refers to the percent of CD3ε⁺, B220⁻, and CD11b⁻ cells found to be positive for Texas Red compared to cells from PBS injected control animals.

Confocal Microscopy.

Cells were incubated for 30 min at 37 °C in 5% CO₂ atmosphere on poly-L-lysine-coated cover slips and then washed twice in PBS. The cells on the cover slips were fixed for 10 min in 3.7% formaldehyde in PBS, blocked and permeabilized with PBS containing 5% normal mouse serum (Santa Cruz Biotechnology, Santa Cruz, CA) and 0.3% Triton X-100, incubated with rabbit anti-CD107A (provided by Professor Minoru Fukuda,³⁰ Burnham Institute for Medical Research) in PBS containing 3% normal mouse serum and 0.1% Triton X-100 and then stained with FITC-labeled goat anti-rabbit antibody in PBS containing 3% normal goat serum (Gibco, Invitrogen Corp., Carlsbad, CA) and 0.1% Triton X-100. The nuclear region was visualized using mounting medium containing DAPI (VECTASHIELD, Vector Laboratories, Inc., Burlingame, CA).

3. RESULTS AND DISCUSSION

3.1. Characterization of PNT-Based Nanoassemblies

Characterization of PEGylated NTs (PNTs).

To generate a cell-type specific and cytoplasmic delivery of PNT-based nanoassemblies in vitro and in intact animals, we first functionalized NTs by ultrasonication and ultracentrifugation of NTs in presence of PL-PEGb (Scheme 1) following a procedure similar to that reported by Kam et al.¹² Pristine NTs are not soluble in aqueous environments and form several microns-long ropes due to the strong van der Waals interactions between the high polarizable sidewalls. Ultrasonication of pristine NTs in presence of PL-PEGb creates local shear that partially divides the end of the ropes in individual NTs (or in aggregates of few NTs) and allows the alkyl chains of the PL-PEGb to hydrophobically adsorb on the isolated ends. The PEG chains of adsorbed PL-PEGb make the ends more hydrophilic which decreases the van der Walls interactions between the NTs of the rope and facilitates the unbundling process. Subsequent ultracentrifugation enriches the solution with short PEGylated individual/very small bundles of NTs (PNTs) that were stably dispersed in PBS in contrast to hydrophobic pristine NTs (Fig. 1(A)). The absorbance spectrum of PNTs in PBS exhibited sharp peaks in the visible range corresponding to the electronic transitions between the van Hove singularities of metallic (M₁₁) and semiconducting (E₁₁ and E₂₂) NTs (data not shown).³¹ This result confirmed that the supernatant fraction, collected after ultrasonication and subsequent ultracentrifugation of NTs in presence of PL-PEGb, was mainly composed by PEGylated individual/very small bundles of NTs. AFM imaging showed that PNTs were short and uniform in length (176 ± 77 nm) (Fig. 1(B)), and had a nonuniform thickness to confirm the successful PEGylation of NTs (Figs. 1(C and D)).

Fig. 1.

AFM of PEGylated carbon nanotubes. (A) Photographs of pristine NTs (left) and of the stably dispersable PNTs (right) in PBS. (B) Size distribution plot of PNTs. The average length and the standard deviation were 176 nm and 77 nm, respectively. (C) 3D-reconstruction of AFM image (dynamic mode, AAC) of PNT deposited on mica. (D) Cross section profile of PNT deposited on mica (inset). The non-uniform thickness confirmed the presence of PEG chains decorating the NT sidewalls.

Loading of Texas Red-Conjugated Neutravin (TRNav) onto PNT-TRNav.

TRNav-decorated PNTs stably dispersed under physiological conditions and were visible by conventional optical and confocal microscopy to confirm the binding of TRNav on PNTs (data not shown). The presence of TRNav on PNTs was further confirmed on the basis of the observation that biotin binding reduced the bandwidth at half height (full-width half-maximum, FWHM) and blue shifted the tryptophan fluorescence emission peak (λ_max) of streptavidin.³² Streptavidin and neutravidin are both tetrameric proteins that have three tryptophans in each monomer. After excitation at 290 nm PNT-TRNav in PBS exhibited an emission band that was blue-shifted and narrower than that shown by free TRNav in PBS (Table I and Fig. 2). We subsequently calculated the number of biotin on PNTs linked to each TRNav. Mixtures having molar ratios between TRNav and biotin from 1:1 to 1:4 were incubated for 2 h before their emission spectra were collected after excitation at 290 nm. The tryptophan fluorescence emission peak was observed to increasingly blue-shift and narrow with increasing biotin (Table I and Fig. 2). PNT-TRNav exhibited a emission band that was blue shifted respect to a TRNav:biotin molar ratio of 1:1 and red-shifted respect to a TRNav:biotin molar ratio of 1:2. The latter results suggest that each TRNav was linked to PNT through approximately 1.5 biotins.

Table I.

Spectroscopic parameters, maximum emission wavelength and full-width half-maximum, for PNT-TRNav dispersed in PBS and for mixtures with different molar ratios of TRNav and biotin in PBS.

Sample	λmaxa (nm)	FWHMb (nm)
PNT-TRNav	339	61
TRNav:biotin = 1:0	345	65
TRNav:biotin = 1:1	342	64
TRNav:biotin = 1:2	337	60
TRNav:biotin = 1:3	334	57
TRNav:biotin = 1:4	332	56

^aMaximum emission wavelength.

^bFull-width half-maximum.

Fig. 2.

Loading of Texas Red-conjugated neutravidin onto PEGylated carbon nanotubes. Normalized emission spectra after excitation at 290 nm of PNT-TRNav dispersed in PBS (black) and for solutions of TRNav and biotin in PBS having molar ratios of 1:0 (red), 1:1 (orange), 1:2 (green), 1:3 (blue), and 1:4 (magenta).

We recorded the absorbance values at 830 nm of several concentrations of PNT-TRNav in PBS, to obtain an extinction coefficient ε_PNT-TRNav(830 nm) of approximately 2.3 × 10⁶ M⁻¹cm⁻¹, and the absorbance values at 283 nm of several concentrations of TRNav in PBS, to obtain an extinction coefficient ${\epsilon }_{\text{TRNav}}\left(283\mathrm{n}\mathrm{m}\right)$ of approximately 1.1 × 10⁵ M⁻¹cm⁻¹. To calculate the number of TRNav per PNT we preliminary obtain the spectrum of TRNav linked to PNTs (A_TRNav@PNT ) by subtracting the absorbance spectrum of PNTs from that of PNT-TRNav matching the absorbance value at 830 nm [A_PNT(830 nm) = A_PNT-TRNav(830 nm)], and then use the equation:

$$N={\mathrm{A}}_{\text{TRNav}@\mathrm{P}\mathrm{N}\mathrm{T}}\left(283\mathrm{n}\mathrm{m}\right)/\left[{\epsilon }_{\text{TRNav}}\left(283\mathrm{n}\mathrm{m}\right)\right.\left.\times L\times c_{\text{PNT-TRNav}}\right]$$

where $N$ was the number of TRNav per PNT, ${\mathrm{A}}_{\text{TRNav}@\mathrm{P}\mathrm{N}\mathrm{T}}\left(283\mathrm{n}\mathrm{m}\right)$ is the absorbance value at 283 nm of TRNav linked to PNTs in PBS, $L$ is the path length and $c_{\text{PNT-TRNav}}$ is the concentration of PNT-TRNav (calculated by A_PNT-TRNav(830 nm) and ε_PNT-TRNav(830 nm)).

This procedure was repeated on 10 batches of PNT-TRNav and we calculated an approximately value of N = 26 ± 3.

Molecular Weight of PNT-Based Nanoassemblies.

We have estimated that a 176 nm long (10, 10) NT has an approximate molecular weight of 3.4 × 10⁵. Since we calculated that 26 ± 3 TRNav (having a molecular weight of approximately 6 × 10⁴) were on each PNT and each macromolecule was linked to 1.5 biotins, we hypothesized that approximately 40 PL-PEGb (having a molecular weight of approximately 3 × 10³) wrapped each NT. Therefore, we calculated the molecular weight of PNTs to be approximately 4.6 × 10⁵ and that of PNT-TRNav to be approximately 2 × 10⁶.

Binding of Antibodies on PNT-TRNav.

PNT-TRNav (40 μg, approximately 2 × 10⁻¹¹ moles of PNTs decorated with approximately 5.2 × 10⁻¹⁰ moles of TRNav, considering 26 TRNav per PNT) in 600 μL of PBS were incubated with 60 μg (4 × 10⁻¹⁰ moles considering a molecular weight of 1.5 × 10⁵) of an equimolar mixture of biotinylated anti-human CD3ε and anti-human CD28 for 1 hour, then washed 3 times in PBS. The molar ratio between the TRNav available on the NT side-walls and the added antibodies was higher than 1 to have complete antibody linkage to the TRNav. We verified this by polyacrylamide gel electrophoresis and western immuno-blotting (Fig. 3). In particular, aliquots of PNT-based nanoassemblies were collected before and after the washes. As shown in Figure 3, both aliquots contained similar amount of antibodies (approximately 0.75 μg) suggesting that all the added antibodies linked to the TRNav decorating the nanoassemblies.

Fig. 3.

Quantitative western blot analysis of the amount of antibodies bound to PNT-TRNav. PNT-TRNav were incubated with an equimolar mixture of biotinylated anti-human CD3ε and anti-human CD28 for 1 hour, and then washed three times in PBS. The first two gel lanes on the left were PNT-based nanoassembly solutions before and after washing three times with PBS. The other lanes were used to run known amount of equimolar mixture of anti-human CD3ε and anti-human CD28 to obtain a calibration curve.

3.2. T-Lymphocytes Membrane Receptor Post-Signaling Endocytosis and Cytoplasmic Delivery In Vitro of PNT-Based Nanoassemblies

To target PNT-based nanoassemblies selectively to T lymphocytes, we decorated it with anti-human CD3ε and anti-human CD28 monoclonal antibodies (PNT-TRNav-α₃α₂₈). Furthermore, to achieve cytoplasmic delivery, we added a synthetic fusogenic polymer (PNT-TRNav-α₃α₂₈p).

Fusogenic polymers are amphipathic compounds that can destabilize the lysosomal membrane by mimicking viruses, which have evolved sophisticated methods to escape from the harsh lysosomal environment and to safely deliver their genetic information into the nucleus. These polymers contain equally spaced acidic and hydrophobic groups. The acidic groups are responsible of the fusogenic activity of the polymer. At neutral pH the acidic groups are deprotonated, leading to electrostatic repulsion that confers a random coil structure to the sequence, whereas are protonated upon a decrease in pH, leading to a decrease in the electrostatic repulsion allowing the polymer to adopt an α-helical structure. In this configuration the polymer can bind the lysosomal membrane and induce vesicle leakage.

The targeting and delivery of nanoassemblies into cells was verified by flow cytometry and confocal microscopy using the Jurkat T leukemia cell line. Upon addition of PNT-TRNav-α₃α₂₈, the internal fluorescence of the cells increased in a time-dependent manner (Fig. 4). The data suggest that binding of PNT-TRNav to CD3ε and CD28 was able to induce signaling through the T cell receptor (TCR), and that signaling-induced endocytosis led to internalization of PNT-TRNav-α₃α₂₈. TCR signaling induced by cross-linked anti-human CD3ε with or without anti-human CD28 is followed by receptor internalization, which is a mechanism of signal downregulation (“post-signaling endocytosis”). Our data suggest that PNT-based nanoassemblies work as cross-linkers to induce TCR signaling. Indeed T cells incubated for 1 h with PNT-TRNav-α₃ or with PNT-TRNav-α₃α₂₈ showed increased expression of CD69 (a marker of T cell activation) when compared to cells incubated with PNT-TRNav (data not shown). Increased signaling following CD28 costimulation led to proportionally increased uptake of nanoassemblies through post-signaling endocytosis. Indeed nanoassemblies with only anti-human CD3ε, but not anti-human CD28, antibodies were internalized at a slower rate and with a different kinetic, suggesting that the costimulation of CD3ε and CD28 membrane receptors by PNT-TRNav-α₃α₂₈ increased the efficiency of the T cell receptor post-signaling endocytotic internalization (Fig. 4). Assemblies without anti-human CD3ε or anti-human CD28 were very weakly internalized by Jurkat T cells, as were assemblies with anti-human CD3ε or anti-human CD28 by Raji B lymphoma cells, which do not express CD3ε or CD28. The observed very low level of non-specific uptake is probably due to residual exposed hydrophobic regions of the NTs.³³

Fig. 4.

T-lymphocytes membrane receptor post-signaling endocytosis in vitro of PNT-based nanoassemblies. Flow cytometry. Fluorescence intensity (A) and uptake efficiency (B) of Jurkat leukemia T cells incubated with PNT-TRNav, PNT-TRNav-α₃, PNT-TRNav-α₃α₂₈, or PNT-TRNav-α₃α₂₈ p and of Raji lymphoma B cells incubated with PNT-TRNav-α₃α₂₈. Untreated cells were used as control. Statistical analysis: one-way analysis of variance (with Bonferroni’s multiple comparison test) for PNT-TRNav-α₃α₂₈ (orange) and PNT-TRNav-α₃α₂₈p (grey) and two-way ANOVA (with Bonferroni’s post-test) for PNT-TRNav-α₃ versus PNT-TRNav-α₃α₂₈ (orange), and PNT-TRNav-α₃α₂₈ p (grey); ** = P < 0.01; * = P < 0.05. Errors bars represent the standard deviation for triplicate experiments

Jurkat T cells readily internalized PNT-TRNav-α₃α₂₈p and exhibited an internal fluorescence that increased in a time-dependent manner and was not statistically significant different from that of cells treated with PNT-TRNav-α₃α₂₈ (Fig. 4). These results suggest that the addition of the fusogenic polymer did not affect the rate of cell entry.

The intracellular location of the nanoassemblies was examined by confocal microscopy. PNT-TRNav-α₃α₂₈ were internalized by cells and transported into the lysosomal compartments (Fig. 5, upper panel) as shown by the yellow fluorescence (images (C) and (D)) corresponding to colocalization of lysosomes (green, image (A) and intracellular nanoassemblies (red, image (B). PNT-based nanoassemblies with the fusogenic polymer were largely cytoplasmic (Fig. 5, lower panel), suggesting that as soon as the endosomes containing the nanoassemblies fuse with the acidic lysosomes, the low pH activates the polymer to disrupt the lysosome membrane, resulting in release of the nanoassemblies into the cytoplasm of the cells.

Fig. 5.

T-lymphocytes cytoplasmic delivery in vitro of PNT-based nanoassemblies. Confocal microscopy. Confocal images of Jurkat T leukemia cells incubated with PNT-TRNav-α₃α₂₈ (upper panel) and with PNT-TRNav-α₃α₂₈ p (lower panel) for 6 hours. Images (A) and (B) showed lysosomal compartments (green) and nanoassemblies (red), respectively. Image (C) is the merge of (A), (B) and nuclear region (blue) while image (D) is the merge of (C) and the bright-field image of the cells. Scale bars approximately 5 μm (upper panels) and 1 μm (lower panel).

The reported results suggest that our PNT-based nanoassemblies could be useful to deliver cargo to the cytoplasm of targeted cells. However, before developing any therapeutic applications short- and long-term cytotoxic effects must be accurately established. Indeed, many nanomaterials (carbon nanotubes,^34–36 fullerenes,³⁷ quantum dots,³⁸ gold nanobeads,³⁹ etc.) have shown signs of toxicity dependent upon such factors as dose, dimension, chemical functionalization, and physical aspects. Nevertheless, the dose of nanoassemblies examined here were not toxic to the cells and did not cause any significant increase in apoptotic or dead cells in the time course considered for trafficking studies (1, 3, or 6 h of treatment), as assessed by staining the cells with annexin V and propidium iodide and analysis by flow cytometry (Fig. 6). The PNT-based nanoassemblies were also not found to be toxic after 24 h of incubation (data not shown).

Fig. 6.

Cytotoxicity of PNT-based nanoassemblies. Percentage of apoptotic (A) and necrotic (B) Jurkat leukemia T cells after 1, 3 and 6 hours of incubation with PNT-TRNav, PNT-TRNav-α₃α₂₈, or PNT-TRNav-α₃α₂₈p. Non-treated cells were used as control. Errors bars represent the standard deviation for triplicate experiments. There were no statistically significant changes with respect to control.

3.3. T-Lymphocytes Membrane Receptor Post-Signaling Endocytosis and Cytoplasmic Delivery In Vivo of PNT-Based Nanoassemblies

We preliminary investigated by flow cytometry the effects of PNT quantity (W) and of the number of antibodies per PNT (R) (molar ratio between antibodies, anti-mouse CD3ε and anti-mouse CD28, and PNTs) on in vivo targeting specificity, measured here as the percent of PNT⁺ cells that fall with in the target cell gate. 129sv/J WT mice were intravenously injected with PBS (control) or PBS solution containing PNT-TRNav-α₃α₂₈ with W and R equal to (W = 3 μg, R = 20), (W = 3 μg, R = 4) and (W = 15 μg, R = 4), respectively. 24 h post-injection, animals were sacrificed and the blood and spleen were assayed for the presence of PNT⁺ cells by flow cytometry (Fig. 7). The results showed that specificity of targeting was mainly dependent on R and not on W. When R was low specificity was poor even increasing W from 3 μg to 15 μg, whereas increasing R from 4 to 20 increased specificity from 7.11% to 55.4% in blood and from 35.5% up to 65.7% in spleen.

Fig. 7.

Effects of PNT quantity (W) and number of antibodies per PNT (R) on specificity of cell targeting in vivo. Total (red blood cell depleted) blood and spleen from control and experimental animals were stained with anti-mouse CD3ε, anti-mouse CD45R (B220), and anti-mouse CD11b and plotted to enumerate T and B lymphocytes. Live PNT⁺ cells (colored events) and total cells (grey) are shown. Numbers within the T cell gate (upper left quadrant) indicate the percent of PNT⁺ cells falling within the gate as a measure of specificity of targeting.

Based on our preliminary findings regarding the effects of PNT quantity and number of antibody per PNT on the specificity, 3 μg of PNT-based nanoassemblies carrying approximately 20 antibodies per PNT were injected intravenously into 129sv/J mice, which were sacrificed and analyzed 24 h later. In animals receiving PNT-TRNav-α₃α₂₈ or PNT-TRNav-α₃α₂₈p, specificity largely favored uptake by CD3ε⁺ T cells (Fig. 8). Therefore the PNT-based nanoassemblies functionalized with anti-mouse CD3ε and anti-mouse CD28 were targeted specifically to T cells in vivo and the addition of the fusogenic polymer did not affect the in vivo cellular uptake.

Fig. 8.

T-lymphocytes membrane receptor post-signaling endocytosis in vivo of PNT-based nanoassemblies. Flow cytometry. 129sv/J WT mice were intravenously injected with PBS (control) or PBS solution containing 3 μg of PNT-TRNav, PNT-TRNav-α₃α₂₈, or PNT-TRNav-α₃α₂₈ p. 24 hours post-injection, animals were sacrificed and the blood and spleen were assayed for the presence of PNT targeted cells by flow cytometry. Total (red blood cell depleted) blood and spleen from control and experimental animals were stained with anti-mouse CD3ε, anti-mouse CD45R (B220), and anti-mouse CD11b and plotted to enumerate T and B lymphocytes. Live PNT⁺ cells (colored events) and total cells (grey) are shown. Numbers within the T cell gate (upper left quadrant) indicate the percent of PNT⁺ cells falling within the gate as a measure of specificity of targeting.

However, even if specificity favored uptake by CD3ε⁺ T cells, fluorescence was seen in CD11b⁺ cells (monocytes and macrophages) and in B220⁺ B cells. We observed approximately 30% to 45% of PNT⁺ cells to be not of the targeted cell type (Fig. 8). Through further flow cytometric analysis we found that in the blood 95% to 99% of the CD3ε⁻ PNT⁺ cells to be CD11b⁺ monocytes and 1% to 5% to be B220⁺ B cells (data not shown). In the spleen we find that 60% to 70% of these CD3ε⁻ PNT⁺ cells are CD11b⁺ macrophages and 30% to 40% of these cells are B220⁺ B cells (data not shown). This result is expected as macrophages express high levels of Fc receptors and are specialized in the clearance of antibody containing complexes and gives further insight into the possible hurdles that will have to be overcome prior to use of this type of material in patients. It should be emphasized that the strategy used here in generating the PNT assemblies could easily be modified for targeted by means other than intact antibodies, such as Fab fragments (which lack Fc regions), peptides, receptor ligands, or aptamers and that this alteration in the targeting strategy could improve specificity. In addition to PNT-TRNav samples shown here our ongoing studies show this effect to be dependent on the presence of targeting antibodies and not the PNT-TRNav (data not shown). Studies targeting other cell types with different antibodies show uptake by CD11b⁺ and B220⁺ cells independent of antibody specificity and that using non-antibody ligand mediated targeting to other cell types does not result in non-specific uptake by these cells (data not shown).

Confocal microscopy of T cells from the animals injected with PNT-TRNav-α₃α₂₈ or PNT-TRNav-α₃α₂₈p confirmed that the nanoassemblies were internalized by the T cells and that the former were in lysosomes, while the latter we also seen outside of these compartments (Fig. 9). Thus, the PNT-based nanoassemblies containing the fusogenic polymer were released into the cytosol of the targeted cells in the animals.

Fig. 9.

T-lymphocytes cytoplasmic delivery in vivo of PNT-based nanoassemblies. Confocal microscopy. Confocal images of T cells from mice intravenously injected with 3 μg of PNT-TRNav-α₃α₂₈ (upper panel) or PNT-TRNav-α₃α₂₈ p (lower panel). Images (A) and (B) showed lysosomal compartments (green) and nanoassemblies (red), respectively. Image (C) is the merge of (A), (B) and nuclear region (blue) while image (D) is the merge of (C) and the bright-field image of the cells. Scale bars approximately 1 μm.

Both in vitro and in vivo, T cells incubated with PNT-TRNav-α₃α₂₈p showed cytoplasmic fluorescence (Figs. 5 and 9, lower panels). Punctate (rather than diffuse) cytosolic fluorescent features corresponding to nanoparticles that had escaped from lysosomes have been previously reported by us⁴ and others^5,17 and could be explained by partial aggregation in the vesicles or, more likely, after their release into the cytosol. However, the cause of this phenomenon is not clear at the moment and further investigation is warranted to understand the effect of surface receptor binding on cytoplasmic nanoassembly aggregation.

4. CONCLUSIONS

In this paper we reported the fabrication of a multivalent, cell-specific and cytoplasmic PNT-based delivery system functionalized with antibodies for cell receptor post-signaling endocytosis and a synthetic fusogenic polymer for disruption of lysosomal compartments. The biomimetic nanoassembly was stably dispersed under physiological conditions and targeted to T lymphocytes both in vitro and in intact animals. The cytoplasmic location of the nanoassemblies was confirmed by confocal microscopy.

Our approach to cell-specific intracellular delivery of a nanoassembly in live animals offers many advantages. NTs are ideal platforms for the construction of multifunctional assemblies and their pharmacodynamic properties can be tuned by varying the phospholipid/PEG coating.²⁹ We also show that addition of a fusogenic polymer to the PNT-based nanoassemblies allows for escape from the lysosomal compartments into the cytoplasm of the targeted cells. This, in turn, opens the possibility of adding intracellularly active proteins or nucleic acids (a “war head”) to the nanoassemblies to achieve a specific cytoplasmic effect in the targeted cell. This may, for example, become useful for treating patients with autoimmune or inflammatory disorders with nanoassemblies armed with apoptosis-inducing proteins specifically targeted into activated myeloid or lymphoid cells. Adding the appropriate hydrolytic enzymes (e.g., cathepsins) as “war heads” to these nanoassemblies may further assist in dissolving the accumulated lysosomal material. It should also be emphasized that PNT assemblies could be targeted by means other than antibodies, such as peptides, receptor ligands, or aptamers. They could also be decorated with viral coat proteins to gain access into the cells targeted by the virus. Armed with antiviral effector molecules, such virus-like nanoassemblies would allow for highly targeted therapies for many currently therapy-resistant viral diseases.