Photothermal therapy (PTT), one of the major modalities for the treatment of cancer, has experienced a tremendous revolution from traditional laser therapy1 to nanotechnology-driven photothermolysis.2, 3, 4 Recently, gold-based plasmonic nanoparticles have shown great promise for use in photothermal cancer therapy.5 The basic rationale is that gold nanospheres (Au NPs) absorb light orders of magnitude stronger than conventional organic dye molecules,6 due to their unique surface plasmon resonance7, 8 (coherent collective oscillation of free electrons of the metal in resonance with the incident light). The strongly absorbed light is converted into heat quickly (on a picosecond time scale) and efficiently (nearly 100% of the absorbed light is converted to heat) by nonradiative processes7, 8, 9, 10 (i.e., electron-electron and electron-phonon interactions). Thus, under electromagnetic irradiation, the nanoparticles generate intense localized heat capable of destroying surrounding targeted cells. These thermal effects require minimal radiation exposure. Current gold analogs under study include spherical Au NPs (Refs. 11, 12, 13, 14, 15, 16, 17, 18), gold nanoshells,2, 19, 20, 21, 22, 23 gold nanorods,4, 24, 25, 26, 27, 28 and gold nanocages.29, 30, 31, 32, 33
To achieve cell- and site-specific delivery for the desired treatment, the nanoparticles must be functionalized with cargoes such as peptides, antibodies, and small molecule ligands for targeting. Arginine-glycine-aspartate (RGD)-peptide-conjugated nanoparticles have been shown to specifically bind to alpha v beta 3 integrin that are overexpressed on some types of cancer cells and tumor vasculature.34, 35 On binding to these receptors, subsequent internalization via receptor-mediated endocytosis leads to intracellular accumulation of the nanoparticles in the cytoplasm. Nuclear targeting of nanoparticles can be particularly challenging, in that there are multiple barriers that exist from the cytoplasmic membrane to the nuclear envelope. To achieve sufficient nuclear targeting, assistance from nuclear translocation peptides, conjugated to the nanoparticles, is required; the nuclear localization signal (NLS) from the simian virus (SV) 40 large T antigen,36, 37 adenovirus,38, 39 and HIV Tat peptide.40
Photothermally induced cell death can occur via apoptosis or necrosis, depending on the laser dosage, type, and irradiation time. It also depends on the subcellular location of the nanoparticles.16, 26 Cell death mechanisms on photothermal treatment have been explored to some degree. For example, the studies of Ton 26 showed that apoptosis is initiated by the disruption of the cytoplasmic membrane. Subsequent influx of calcium ions induces membrane blebbing and damage of actin filaments. In the work by Pustovalov, 16 laser pulses could induce cell damage by a series of photothermal and accompanied phenomena: denaturation and breakdown of proteins, cell cavitation, cellular structure rupturing, evaporation of cellular liquid and bubble formation by shock waves due to particle thermal expansion, and optical breakdown or plasma generation of Au NPs.
In this paper, we conduct a comparative study of the photothermal damage of cancer cells using continuous wave (cw) or nanosecond-pulsed lasers with Au NPs located at the nucleus or in the cytoplasm of the cell. Au NPs are conjugated with NLS (GGGPKKKRKVGG) peptides originating from the SV 40 large T antigen for nuclear localization and linear RGD peptides (RGDRGDRGDRGDPGC) for cytoplasm localization. The energy threshold and cell death pathways are determined and described according to the types of lasers and the location of Au NPs within the cell.
Materials and Methods
Synthesis of Au NPs and Peptide Conjugation
Citrate-capped Au NPs, in diameter, were synthesized according to the citrate reduction method.41, 42 Briefly, of a 0.01% (by weight) auric acid (HAuCl4, Sigma-Aldrich) aqueous solution was heated to boiling while stirring in a beaker. A volume of a 1% (by weight) trisodium citrate (Sigma-Aldrich) aqueous solution was quickly added and the solution changed color within several minutes (yellow to black to red), indicating nanoparticle formation. The surface plasmon absorption spectrum of the nanoparticle solution was measured with a UV-Vis spectrometer (Ocean Optics). Nanoparticle size was determined by transmission electron microscopy (TEM) and dynamic light scattering (DLS) using the Zeta Potential Instrument ZetaSizer Nano ZS90. Prior to peptide conjugation, Au NPs were modified with poly(ethylene) glycol (PEG). This was done by mixing citrate-capped Au NPs (concentration calculated using as determined by Hurst 43) with mPEG-SH 5000 (methoxy-PEG, MW 5000, Laysan Bio Inc.) for at room temperature. The PEGylated nanoparticles were centrifuged at for and redispersed in phosphate-buffered saline (PBS) to a final optical density of 2.0. Citrate-capped Au NPs exhibit a zeta potential around . After PEGylation, the zeta potential of Au NPs decreases to . Having a nearly neutral charge, the PEG serves to prevent nonspecific adsorption of proteins to the nanoparticles in a physiological environment. To maximize the number of peptides that will bind to the nanoparticles, the amount of PEG molecules on the nanoparticles should be minimized. Here, the loading ratio of PEG molecules to Au NPs is PEG molecules per nanoparticle, which is much lower than the number of PEG molecules on a gold surface for a densely packed monolayer of Au NPs (as estimated by an SPDP assay). With minimal PEG molecules on the Au NP surface, the thiolated peptides are able to bind to unoccupied sites on the Au NPs via Au-S bonds.
To facilitate binding to Au NPs, NLS peptides were functionalized with thiol linkers, as described in previous work,36 and was modified with a cysteine residue. The thiol-modified NLS or RGD was mixed with PEGylated Au NPs to a final peptide concentration of and allowed to react overnight. The peptide-conjugated Au NPs were purified by centrifugation at for and redispersed in PBS buffer.
Cell Culture and Cellular Incubation with Au NPs
Human oral squamous cell carcinoma (HSC-3) cells were cultured in Dulbecco’s modification of Eagle’s medium (DMEM, Cellgro) with 10% fetal bovine serum (FBS, Gem Cell) at under 5% . The cells were cleaved from the culture dish by trypsin and replated onto glass coverslips in a 12-well tissue culture plate to grow for . The coverslips were coated with collagen type I (Roche), in advance, for optimum cell growth. Peptide-conjugated Au NPs [optical density at ] were then added to each well at a volume of and cells were returned to under 5% for . The final concentration of Au NPs in cell culture medium was , as previously reported for PTT using Au NPs in vitro.12 After incubation, the cells on the glass coverslips were washed with PBS to remove excess Au NPs before further treatments.
In this study, two types of lasers were used for photothermal therapy: a cw argon ion laser with a wavelength and a -switched Nd:YAG nanosecond-pulsed laser with a wavelength. Note that the absorption peak of gold nanospheres is at , the OD at is about 94.5% of the OD at , thus, the difference in laser wavelength here might not affect the difference in the energy threshold between the cw and pulsed lasers. The single pulse of the nanosecond laser was achieved by the -switched control, and the pulse width was about . For the irradiation experiment, both the cw and pulsed lasers were focused to in diameter at the sample position. Multiple regions of the cell monolayer were exposed to the cw laser beam at different energy densities for each and to a single nanosecond laser pulse with different energy. All experiments are repeated at least 3 times for energy threshold determinations to confirm reproducibility. Immediately following laser treatment, the cells are stained with 0.4% trypan blue (Sigma) for . Dead cells were stained blue due to the accumulation of the dye, while live cells remained unstained. For easy comparison, the total energy input is used to compare the energy threshold between the cw and pulsed lasers.
Fluorescence Staining and Imaging
To investigate the pathways of cell death on laser treatment, a double-stain (molecular probe) was used to distinguish between cell necrosis and apoptosis. The apoptotic cells were stained with annexin V-Alexa Fluor 488 (AV), while necrotic cells were stained with propidium iodide (PI). A volume of AV and a volume of PI were added to cells in of annexin-binding buffer and incubated at room temperature for . After staining, cells were washed with PBS and fixed with a 4% paraformaldehyde solution. Fluorescence images were taken using an Olympus IX 70 microscope equipped with an argon ion laser. The excitation wavelength was and the laser power was .
Results and Discussion
Selective Intracellular Delivery of Peptide-Conjugated Au NPs
Au NPs were modified with NLS and linear peptides for cell nuclei and cytoplasm targeting, respectively. The conjugation was done by binding thiolated NLS and cysteine-modified RGD peptides to Au NPs that were stabilized with a minimal amount of thiolated methyl-poly (ethylene) glycol 5000 (mPEG-SH 5000) in advance. The pre-PEGylation serves two functions: (1) to decrease the surface charge of the citrate-capped Au NPs so as to avoid particle aggregation when the nanoparticles are exposed to oppositely charged peptides and (2) to minimize particle-cell nonspecific interactions, as citrate-capped Au NPs have shown to be internalized by both normal and cancer cells with high efficiency.44
Figure 1 shows dark-field light scattering images of HSC-3 cells after a incubation with NLS-modified Au NPs (NLS/Au NPs) [Fig. 1] and RGD-modified Au NPs (RGD/Au NPs) [Fig. 1]. While NLS/Au NPs were dominantly localized at the cell nucleus, the RGD/Au NPs were distributed throughout the cytoplasm. This phenomenon was seen in over 80% of cells for several trials of this study. These observations can be attributed to the ability of the virus mimetic NLS to deliver nanocarriers to the cell nucleus via its binding to importin proteins on the nuclear envelope45 and the ability of RGD to enable intracellular incorporation of particles via receptor-mediated endocytosis on binding to integrin receptors on the cell membrane of cancer cells.46
Energy Threshold of Photothermal Therapy
Au NPs localized at the nucleus
For in vivo and clinical applications, it is favorable to use a minimal amount of energy in PTT so as not to damage surrounding healthy tissue. In this study, we determined the energy threshold required for cell destruction, utilizing trypan blue staining for detection of dead cells. Figure 2 displays bright-field images of cells (incubated with NLS/Au NPs) after single-pulse nanosecond laser treatment at different energies. It is apparent here that a single pulse (pulse duration: ) with an energy threshold of can induce cell death (Fig. 2, top right). In the control experiment (data not shown), in which the cells were not treated with Au NPs, a single pulse with a minimum energy of was required to induce cell death. This is approximately 10 times higher than that required for NLS/Au-NP-treated cells. Such improvement in the photothermal efficiency by treatment with NLS/Au NPs is undoubtedly due to the laser heating of Au NPs localized at the nucleus. The wavelength of the nanosecond laser overlaps with the surface plasmon absorption band of the Au NPs (absorption maximum around ). The surface plasmon absorption of the laser light by the Au NPs ensures efficient light energy conversion to thermal energy, which induces cell death via hyperthermia or cell ablation.
A cw argon ion laser with a wavelength of and laser energy of was used to treat cells with Au NPs localized at the nucleus. We did not observe any dead cells under these conditions (images not shown). This laser energy is almost 4 times higher than the laser energy required to destroy cells with Au NPs localized on the cytoplasmic membrane12 . Such a difference clearly shows that the location of Au NPs affects the photothermal efficiency to a great degree.
Au NPs located in the cytoplasm
Figure 3 displays bright-field images of cells (incubated with RGD/Au NPs) after treatment with pulsed nanosecond [Fig. 3] and cw [Fig. 3] lasers. In the case of the nanosecond-pulsed laser treatment, the energy threshold (minimum energy at which cell death occurs) was [Fig. 3, bottom left] with a single pulse (pulse duration: ), which is almost 3 times the energy threshold for cells with Au NPs localized at the nucleus. With this observation, it is apparent that the location of Au NPs, either at the nucleus or throughout the cytoplasm, determines the laser energy required to induce cell death. As shown in the dark-field light scattering images in Fig. 1, Au NPs are more localized in cells with NLS/Au NPs, but more dispersed in cells with RGD/Au NPs. Note that in these experiments, the same concentration of peptide conjugated Au NPs was used for the two cases and all other parameters were controlled. Therefore, the close packing (aggregation) of Au NPs at the nucleus must generate heat more efficiently, providing for a lower energy threshold. This might be due to two effects. First, an increased surface plasmon field is expected due to the aggregation of Au NPs at the nucleus. Second, the high laser intensity (due to short pulse width) enhances nonlinear absorption processes. These two effects make the energy absorbed during the pulse so high that it requires a lower threshold to destroy the cell.
In contrast to the nanosecond-pulsed laser, a cw laser requires a much lower energy to induce cell death when Au NPs are in the cytoplasm compared to Au NPs at the nucleus. As seen in Fig. 3, at [Fig. 3, middle], cell damage is achieved with Au NPs in the cytoplasm, whereas did not induce cell destruction with Au NPs localized at the nucleus (data not shown), as we described in the previous section. This contrast could be due to the difference in the amounts of effective energy delivered to the cell from the two different lasers. This depends on the laser power and plasmon field intensity, which depends on the degree of nanoparticle aggregation. These factors may also dictate the cell death mechanisms involved when using the different types of lasers.
In summary, the energy threshold for the pulsed nanosecond laser is about 500,000 times lower than that for the cw laser when Au NPs are localized at the nucleus, and about 100,000 times lower when Au NPs are dispersed throughout the cytoplasm. In general, when Au NPs are localized at the nucleus, a lower energy nanosecond laser pulse is required to induce cell death than when Au NPs are dispersed throughout the cytoplasm, but the opposite is true when the cw laser is used. Previous results12 showed that when Au NPs are localized on the cell membrane, the intensity threshold required to induce cell death with the cw laser is , which is lower than that required when Au NPs are in the cytoplasm , as determined here. Tong also reported similar results using folate-targeted gold nanorods.26 Such a phenomenon could be due to the arrangement and the density of the nanoparticles on the cell membrane and the potential for aggregation of Au NPs, resulting in a stronger surface plasmon absorption of light and thus more efficient photothermal effects.
Photoacoustic effects could potentially be the cause of cell death for the nanosecond-pulsed laser treatments displayed here. On irradiation with a laser pulse, the compact Au NPs at the nucleus could produce highly localized heat, generating acoustic waves.16 This could cause mechanical damage to the cell via the formation of bubbles inside cells. Multiphoton processes could also take place, leading to the dissociation and ionization of nucleic acid bases, thereby destroying DNA or protein molecules.
Cell Death Pathways
Au NPs localized at the nucleus
Because laser-induced cell death was observed, as already discussed, it is important to investigate the mechanisms by which cell death occurs. To do so, a standard apoptosis kit was used. The kit contains a probe for cells that exhibit early apoptosis (AlexaFluor 488 AV) and a probe for cells that are necrotic (PI). Figure 4 shows bright-field images after trypan blue staining and fluorescence images after AV/PI staining of the cw [Fig. 4] and nanosecond pulsed laser [Fig. 4] treated cells with Au NPs localized at the nucleus. As stated in the previous section, cell death does not occur (no accumulation of trypan blue in cells) even at a cw laser energy of [Fig. 4, bright-field image], indicating that the cell membrane is intact during cw laser treatment. However, the cells in this case display AV/PI fluorescence when is used [Fig. 4]. Immediately following cw laser treatment, the cells were stained green by AV, indicating an early stage of apoptosis. Cell membrane alterations can occur during the initial stages of apoptosis, namely the translocation of phosphatidylserine (PS) (normally located on the cytoplasmic side of the cell membrane) to the external surface of the cell. The -dependent phospholipid-binding protein, AV, has a high affinity for PS and can thus probe for early apoptosis. At later times following cw laser treatment, a time-dependent PI staining of the cells occurs. PI is a DNA-binding dye that is not cell membrane permeable and thus distinguishes necrotic cells from apoptotic cells. From the fluorescence images at different time points following laser treatment [Fig. 4, middle and bottom] the progression of cells from early apoptosis to necrosis is apparent as the fluorescence changes from green (AV), to yellow (AV and PI), and finally to red (PI). (Color online only.)
Apoptosis is not observed when a nanosecond-pulsed laser is used for the treatment of cells with Au NPs localized at the nucleus. At a pulsed laser energy of (below the threshold of ), no apoptosis was observed. On increasing the pulsed laser energy to , only PI (red fluorescence) accumulation in cells is observed [Fig. 4, right], indicating that the higher energy nanosecond laser pulse induced cell death via necrosis. This immediate cell death is due to the strong and rapid heating of Au NPs by the nanosecond laser pulse, which could produce shock waves16 that damage the cells in a very short time. The absence of green fluorescence from AV staining, in this case, might be due to the nanosecond laser pulse inducing severe damage to the cell membrane, which would make the binding of AV to PS difficult.
Au NPs located in the cytoplasm
Figure 5 shows bright-field images after trypan blue staining and fluorescence images after AV/PI staining of the cw [Fig. 5] and nanosecond-pulsed laser [Fig. 5] treated cells with Au NPs dispersed throughout the cytoplasm. When the cw laser energy was at (below the threshold of ), cells were not stained by trypan blue, even with a laser irradiation, indicating that the laser treatment did not cause cell membrane damage. The fluorescence images at different time points after the laser treatment showed the progression of cells from early- to late-stage apoptosis (not reaching complete necrosis). When the cw laser energy was higher, [Fig. 5, bottom], the dead cells are stained by both trypan blue and PI, indicating cell membrane damage and resulting in necrosis of the cells. The cell death mechanism’s dependence on laser energy, as seen here, was also observed by other researchers28 when folate-targeted gold nanorods were delivered to the cytoplasm of cancer cells.
Similar to the results obtained when Au NPs are localized at the nucleus, the nanosecond-pulsed laser immediately induced necrosis in cells with Au NPs distributed throughout the cytoplasm, as indicated by the accumulation of PI in the cells [Fig. 5, right]. It is apparent, with this observation, that the location of the Au NPs within the cell does not determine the mechanism of cell death on nanosecond-pulsed laser treatments. However, when the cw laser is used, higher energy induces necrosis and lower energy induces apoptosis when Au NPs are located in cytoplasm, while no necrosis is observed when Au NPs are localized at the nucleus. This is most likely due to there not being sufficient laser energy for a necrosis process to occur.
In summary, Au NPs can selectively target the nucleus or the cytoplasm of HSC-3 (cancer) cells using NLS and linear RGD peptides, respectively. Photothermal studies show that the nanosecond-pulsed laser is more effective than the cw laser for purposes of inducing cell destruction, independent of the intracelluar location of the Au NPs. Photothermal experiments also show that when treating cells with the cw laser, Au NPs are more successful at destroying cells when distributed in the cytoplasm than when localized at the nucleus, while the opposite is true with nanosecond-pulsed laser treatments. Fluorescence staining results indicate that when cells are treated with a cw laser, cell death is attributed to apoptosis with Au NPs localized at the nucleus and apoptosis or necrosis, depending on the laser intensity, with Au NPs in the cytoplasm. The nanosecond-pulsed laser treatments induce necrosis only, despite the location of the nanoparticles.
X. Huang would like to thank the distinguished CCNE postdoctoral fellowship at Emory-Georgia Tech Cancer Center for Nanotechnology Excellence, B. Kang thanks the support of China Scholarship Council (No. 2008683010) and Doctor Innovation Funds of NUAA (BCXJ08-09, China). I. H. El-Sayed thanks the Hearing Research Institute for financial support and M. A. El-Sayed thanks the honor of Julius Brown Chair at Georgia Tech.