COMMUNICATION

DOI: 10.1002/adma.200800941

Ultra-Low Energy Threshold for Cancer Photothermal Therapy Using Transferrin-Conjugated Gold Nanorods**
By Jing Liang Li,* Daniel Day, and Min Gu*
Gold nanoparticles have shown potential applications for cancer detection and localized photothermal therapy (PTT).[15] Among the gold nanoparticles that have been studied, gold nanorods are of particular interest because of their tuneable near-infrared (NIR) absorption and the recent success in their size-controlled, large-scale synthesis.[6,7] The strong two-photon photoluminescence (TPL) of gold nanorods render them good contrast agents for cancer cell imaging under two-photon excitation,[8,9] which makes it more suitable to three-dimensional nonlinear optical imaging of biological samples.[10,11] Despite the successful application of gold nanorods in PTT and in vitro cancer imaging, this technique still has a long way to go before it can be implemented clinically. For clinical applications, the energy input should be as low as possible to avoid damage to healthy tissues, which is always a concern for laser-based applications. The energy threshold can be reduced by optimizing tumor targeting of the particles, choosing a suitable laser mode, and improving the light absorption efciency of the nanorods. It has been observed that a pulsed femtosecond laser beam in PTT was more effective than a continuous wave (CW) laser beam.[12,13] However, most of the gold nanorods are randomly oriented in cells, making it impossible to achieve the maximum energy efciency due to the limited fraction of excitation of the total nanorods under linearly polarized light. To increase the light absorption efciency and hence the photothermal effect of the nanorods, a circularly polarized laser beam, which can emit a light beam of all polarization angles within one optical period, can be used to activate as many nanorods as possible, as illustrated by Figure 1a. To demonstrate the circular polarization effect on nanorod excitation, nanorods from a dilute solution were deposited on a cover slide and the same population of nanorods was subject to a linearly and circularly polarized light illumination in sequence. The corresponding images of the nanorods are given in Figure 1b and 1c, respectively. Both single gold nanorods and aggregated nanorods (larger and brighter spots as circled in Fig. 1b) can be identied. It shows that under illumination of circularly polarized light at the same incident power, more nanorods can be excited and be imaged clearly. To demonstrate the circular polarization effect for cancer therapy, we have developed biofunctional nanorods for the specic targeting of cancer cells. Transferrin has been shown to be effective as a ligand for actively targeting malignant cells for the delivery of drugs, proteins and genes.[14,15] Although it has been used to enhance the cellular uptake of gold nanoparticles[16] and quantum dots/rods,[17] conjugating transferrin to gold nanrods for combined imaging and therapy of cancer cells has not been reported. Our work shows the successful development of transferrin-conjugated gold nanorods that can be used for efcient targeting and imaging of cancer cells. More interestingly, the use of circularly polarized femtosecond light for two-photon-induced cancer PTT with gold nanorods lowers the damage energy threshold to one order of magnitude below the medical laser safety standard, making it potentially viable for medically safe applications. In this work, an inverted scanning optical microscope coupled with a femtosecond laser beam (see Experimental section for details) was used for simultaneous TPL imaging and PTT of human cervical cancer cells, HeLa, at a wavelength of 800 nm. A quarter-wave plate was used to convert the linearly polarized beam into a circularly polarized beam to increase the light absorption efciency of the nanorods. To increase the cellular uptake of gold nanorods, the surface of gold nanorods was conjugated with transferrin, which can target the transferrin receptors on HeLa cells. To stabilize the conjugated nanorods in the culture medium, the surface of the nanorods was also conjugated with polyethylene glycol (PEG). The structure of a conjugated gold nanorod with a mixed layer of PEG and transferrin molecules is schematically shown in Figure 2a. Figure 2b is a transmission electron microscopy (TEM) image of the as-synthesized gold nanorods with an average length of 45 nm and an aspect ratio of 4. These surfactant cetyltrimethylammonium bromide (CTAB)-coated nanorods (CTAB-NRs) display a maximum longitudinal absorbance at a wavelength of 787 nm in water (Fig. 2c). When dispersed in the culture medium RPMI, the spectra became broader and the maximum absorbance wavelength shifted to 795 nm. This indicates that slight aggregation might occur because of a decrease in surfactant coverage on the nanorods. Precipitation occurred after the solution was left static at room temperature for a few hours. The PEG-conjugated (PEG-NRs) and transferrin-conjugated nanorods (PEG-TfNRs) showed good stability in the culture medium, as indicated by their absorbance spectra. The maximum absorption of the

[*] Dr. J. L. Li, Prof. M. Gu, Dr. D. Day Center for Micro-Photonics Faculty of Engineering and Industrial Sciences Swinburne University of Technology Hawthorn, VIC 3122 (Australia) E-mail: jili@swin.edu.au; mgu@swin.edu.au [**] The authors made equal contributions to this work. The authors would like to thank Mr. Peter Zijlstra for help with the TEM measurements and the Australia Research Council for nancial support.

3866

2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2008, 20, 38663871

COMMUNICATION
3867

TPL, and cell transmission images after being incubated with the nanorods for 4 and 6 h, respectively. After six hours of incubation, the cellular uptake of the nanorods was so signicant that the shape of the cells could be identied. In contrast, due to the nonspecic binding of PEG to cells, the cellular uptake of PEG-NRs nanorods was not signicant even after six hours of incubation (Fig. 3c). The pronounced difference of the cellular uptakes between PEG-NRs and PEG-Tf- NRs indicates that the chemical surface properties of the nanorods play a signicant role in controlling the uptake of the nanorods by the cells. When the nanorods were conjugated with transferrin molecules, the particles display a much greater afnity to the cells due to the interaction of the transferrin molecules with the transferrin receptors on the membrane of the cells. The transferrin receptor on HeLa cells has been well characterized and it has been reported that that transferrin receptor was overexpressed on HeLa cells.[18] The transferrin-mediated uptake of gold nanorods was successfully demonstrated in a recent work,[19] in which transferrin molecules were physically attached to the surface of Figure 1. a) A schematic description of enhanced plasmon absorption of gold nanorods in cells gold nanorods for dark-eld scattering by converting linearly polarized light into circularly polarized light. The circularly polarized light can imaging of HeLa cells. To further prove activate a larger number of randomly oriented nanorods (red) in cells than the linearly polarized that the uptake of nanorods by the cells is beam along the vertical direction. b) Image of nanorods excited by linearly polarized light and primarily mediated by the transferrinc) image of the same nanorods excited by a circularly polarized light. The circled area in (b) indicates aggregated gold nanorods. To prevent the possible damage of the nanorods by transferrin receptor interaction, the cells repeated scanning, the incident power for excitation was kept at a very low value of 100 mW for were preblocked with free transferrin moleboth linearly and circularly polarized illumination. The scale bars are 5 mm. cules for one hour before the PEG-Tf-NRs were supplied. The results showed that the cellular uptake was signicantly reduced due to the presence of the blocking protein-coated gold nanorod solution shifted to a wavelength molecules (Fig. 3d), which proved the importance of the of 790 nm due to the change in refractive index of the particle transferrin-transferrin receptor interaction in controlling the surface. Slight precipitation did occur in the solution of the cellular uptake of the nanorods. The uptake of PEG-Tf-NRs by PEG-NRs and PEG-Tf-NRs. However, the particles could be blocked cells is similar to that of PEG-NRs (Fig. 3c) because of redispersed simply by hand shaking, which was in contrast to the non-specic uptake in these two cases. Due to the abundant the irreversible aggregation of the CTAB-NRs. Broadening of transferrin receptors on malignant cells, the transferrinthe spectra was also observed for PEG-Tf-NRs, which transferrin receptor interaction should be an effective pathway indicates that the size distribution of the particles might for enhancing the cellular uptake of nanoparticles. The above become broader in the presence of a protein coating. We results also indicate that two-photon excitation is a good observed that the PEG-NRs and PEG-Tf-NRs could be used technique for in vitro imaging of gold nanorods and thus for after long storage periods, i.e., of a few months, without causing cancer diagnostics as long as the nanorods are functionalized obvious changes in their cellular uptake. Figure 2d shows the with suitable targeting molecules. Due to the severe aggregation dependence of the TPL intensity I on the laser excitation and precipitation of the CTAB-coated gold nanorods in culture power P on a double logarithmic scale; the slope of the curve medium, negligible uptake of these particles by cells was was 2.04, indicating that the photoluminescence from the gold observed (image not shown). nanorods resulted from two-photon excitation. Figure 4 shows the PTT results. Ethidium bromide (EB) was Figure 3a and 3b show the combined TPL images of PEGused to stain the nuclei of dead cells. EB is membraneTf-NRs nanorods, transmission images of cells and combined impermeable to live cells and thus can be used to examine the

Adv. Mater. 2008, 20, 38663871

2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advmat.de

COMMUNICATION
3868

membrane integrity of cells. It was observed that in the absence of gold nanorods, the cells could not be killed at a laser power of 20 mW after 60 scans (1.05 s per scan, scanning area 60 mm 60 mm) (Fig. 4a) (linear polarization). Clear cell death was observed at 25 mW after 20 scans (Fig. 4b). In the absence of gold nanorods, the thermal effect of the laser on cells should not be signicant due to the low absorption of the cells of NIR light. However, the highly focused laser could induce the disruption of membrane structures, compromising membrane integrity.[20,21] However, in the presence of nanorods, the laser power effective for therapy was observed to be 0.5 mW, indicating that the photothermal effect of nanorods contributes largely to the cell death. Figure 4c shows the cells after being scanned 150 times (threshold). Interestingly, when the linearly polarized light was converted into circularly polarized Figure 2. The structure and properties of biofunctionalized gold nanorods. a) The structure of a PEG-Tf-NR. b) TEM image of nanorods. c) UV-visible absorption of CTAB-NRs in water, PEG-NRs and light, the number of scans at the same PEG-Tf-NRs in culture medium RPMI. d) Dependence of uorescence intensity of gold nanorods on power needed to damage cells was excitation power. reduced to 30, indicating the higher efciency of the nanorods under circularly polarized illumination (Fig. 4d). When the laser power was increased to 1.0 mW, the cells were killed after 40 and 10 scans under linearly and circularly polarized light, respectively (Fig. 4e and 4f). The minimum number of scans required as a function of the incident power is shown in Figure 4g. The results indicate that the scanning duration could be reduced by converting the linearly polarized light into the circularly polarized light. For example, when the laser power was 1.5 mW, cells could be killed after 10 and 5 scans for linearly and circularly polarized beams, respectively. A lower power of 0.2 mW under circular polarization illumination was also effective after 150 scans. No cell death was observed at 0.1 mW after 400 scans. This means that 0.2 mW might be the threshold power for the circularly polarized light illumination, which is two orders of magnitude lower than the threshold observed in the absence of gold nanorods. The reduced exposure time required for circularly polarized light indicates Figure 3. Combined TPL of nanorods and transmission images of HeLa cells incubated with gold that this illumination method is more nanorods: a) PEG-Tf-NRs, 4 h; b) PEG-Tf-NRs, 6 h; c) PEG-NRs, 6 h; and d) PEG-Tf-NRs, 6 h, energy-efcient compared to the linearly blocked with free transferrin molecules. Scale bars: 10 mm.

www.advmat.de

2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2008, 20, 38663871

COMMUNICATION
3869

polarized light for the laser power from 1.5 mW to 0.5 mW. The above results indicate that a longer laser exposure and higher energy uence are needed when the laser power is reduced, which means that a higher power is more effective to destruct cells. When a laser beam operating at a high power is irradiated on cells labeled with gold nanorods, heat could be built up quickly in the nanorods and released to cells. This quick energy transfer could destroy the cell membrane integrity more efciently. This is veried by the more intense membrane blebbing shown in Figure 4e and f at the higher laser power. Membrane blebbing has been proven a mechanism of cell death in the presence of gold nanorods.[13] The thresholds of energy densities observed in this work are comparable to those reported in the literature when CW-mode laser was used.[1,4,23] However, in this work, the effective illumination time per scan for nanorods is only about 38.0 ms (see experimental section for calculation). Therefore, the total illumination time ranges from 0.19 (5 scans) to 5.70 (150 scans) milliseconds, signicantly less than the exposure durations of a few minutes reported in literature. At the same level of the power density, these short exposures result in energy uence a few orders of magnitude lower. In case of a femtosecond laser, the energy can be more efciently conned by Figure 4. Photothermal therapy of cancer cells in the absence and presence of gold nanorods: gold nanoparticles.[12] This together with the a) no nanorods, 20 mW, 60 scans; b) no nanorods, 25 mW, 20 scans; c) with PEG-Tf-NRs, linear polarization (LP), 0.5 mW, 150 scans; d) with PEG-Tf-NRs, circular polarization (CP), 0.5 mW, efcient targeting of the cancer cells by the 30 scans; e) with PEG-Tf-NRs, LP, 1.0 mW, 40 scans; f) with PEG-Tf-NRs, CP, 1.0 mW, 10 scans. particles contribute to the low energy level Scale bars: 10 mm. g) Dependence of the minimum number of scans on the incident mean power required for cell therapy. With the implemenof linearly and circularly polarized light. tation of circularly polarized light, the light absorption and conversion by the gold nanorods were further improved. polarized light at the same power. The energy uences were In summary, this work reports the efciency of transferrincalculated and given in Table 1. Under either linear or circular conjugated gold nanorods in targeting, two-photon imaging and polarization illumination, the energy uences at the laser photothermal therapy of HeLa cancer cells. The results powers used in this work are below the medical safety level of showed that the cancer cells can be clearly imaged after a 100 mJ cm2.[22] The circularly polarized light reduces the few hours incubation with the nanorods. The signicant uence from half to one fth of those required for linearly cellular uptake could be attributed to the transferrintransferrin receptor interaction. Due to the general upregulaTable 1. Energy uence for cell therapy. tion of transferrin receptors on cancer cells, the biofunctional gold nanorods can potentially be used for the imaging of a Mean Power Power density Energy uence number of cancer cells. Under both linearly and circularly 2 2 [mW] [W cm ] [mJ cm ] polarized femtosecond illumination, the energy thresholds for linear circular cell therapy were observed to fall below the medical safety level. The low energy thresholds are attributable to the high cellular 1.5 41.7 15.8 7.9 uptake of the gold nanorods. A circularly polarized light was 1.0 27.8 31.7 10.6 0.7 19.4 59.1 11.1 found to be more efcient in cancer cell therapy. By converting a 0.5 13.9 79.2 15.8 linearly polarized light into a circularly polarized light, the 0.2 5.6 ineffective 31.7 energy threshold was signicantly reduced, making the photo-

Adv. Mater. 2008, 20, 38663871

2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advmat.de

COMMUNICATION

thermal therapy medically safer. Due to the large difference in the threshold power for unlabeled cells and gold nanorodtargeted cells, simultaneous imaging and highly localized PTT can be achieved without harming the healthy tissue exposed to a laser beam.

Experimental
Materials: Gold(III) chloride trihydrate (!99.9%), sodium borohydride (99.99%), L-ascorbic acid (!99.0%), CTAB (!99.0%), O-(2-Aminoethyl)-O-Methylpolyethylene Glycol 5000 (NH3-PEG, MW 5000), transferrin, N-ethyl-N0 -(3-dimethylaminopropyl) carbodiimide hydrochloride (EDAC) (!99.0%) and N-Hydroxysuccinimide (NHS) (98%) were obtained from Sigma. All the chemicals were used as received. A cancerous cell line HeLa (cervical cancer) was obtained from American Type Culture Collection (ATCC). Nanorods Preparation: Gold nanorods were prepared following a reported method, but scaled up to 400 mL. [6] Briey, gold seeds were prepared by adding freshly prepared ice-cold NaBH (10 mM, 0.6 mL) into a solution containing CTAB (100 mM, 7.5 mL) and HAuCl4 3H2O (10 mM, 0.25 mL), which were mixed for 2 minutes. The color of the seed solution was pale brown. To prepared gold nanorods, seed solution (2 mL) was added into a nanorods growth solution, mixed gently and left still for ten hours. The growth solution was prepared by adding the following reagents into a 500 mL conical glass ask in the following order with gentle mixing: CTAB (100 mM, 400 mL), HAuCl4 3H2O (10 mM, 17 mL), AgNO3 (10 mM, 2.5 mL), and ascorbic acid (100 mM, 2.7 mL). Bioconjugation of Gold Nanorods: The conjugation of gold nanorods with PEG and transferrin, the CTAB molecules on the surface of gold nanorods were replaced with thioglycolic acid molecules. To link the particles with thioglycolic acid, the nanorods were centrifuged and washed twice to reduce the CTAB concentration and dispersed in 2-(N-morpholino)ethanesulfonic acid (MES) buffer (pH 5.5, 10 mL). Then, thioglycolic acid aqueous solution (1 mM, 50 mL) was added and the solution was magnetically stirred for three hours. Then, EDAC (100 mM, 100 mL) and NHS (100 mM, 250 mL) solutions were added into the glycolic acid-coated nanorod solution. The carboxylic groups on the particle surface were activated to form reactive NHS ester intermediates. After 30 min, the activated particle solution was centrifuged and the particles were dispersed in phosphate buffered saline (PBS) solution (pH 7.5, 10 mL) containing NH3-PEG (1.5 mM) and transferrin (0.65 mM). The solution was stirred for another two hours. The amine groups on the protein and PEG molecules reacted with the active ester groups on the surface of the gold nanorods to form stable amide bonds. The excess reactants and by-products were removed by centrifuge. The conjugated nanorods were dispersed in RPMI culture medium for cell culture. Cell Culture: The cells were cultured in a RPMI medium supplied with fetal bovine serum (10%) at 37 8C under 10% CO2 atmosphere. For two-photon imaging and therapy of the cells, the cells were incubated with the gold nanorods for a desired time in 24-well culture plates with thin glass cover slides on the bottom of the wells. Cells incubated with PEG-coated gold nanorods were used as controls. To verify that the cellular uptake was primarily mediated by interactions between transferrin and transferrin receptors, the cells were preblocked with free transferrin molecules for 1.5 h. The concentration of the free transferrin molecules was a thousand times of that for conjugation. Two-Photon in-vitro Imaging and Photothermal Therapy: The in vitro two-photon imaging was carried out on a Fluoview inverted scanning microscope (FV300). A femtosecond Ti:sapphire laser (MaiTai, Spectra Physics) with a repetition rate of 80 MHz was used for the two-photon excitation and photothermal therapy of cancer cells. The wavelength was xed at 800 nm and a water immersion 60

objective lens (N.A. 1.20) was used. For imaging nanorods on a cover slide, the surfactant CTAB-coated nanorods were centrifuged, washed three times, and redispersed in water. A small volume of dilute nanorod solution was deposited on a cover slide. The sample was dried naturally under room temperature. The same area of the slide was illuminated by a linearly and circularly polarized light in sequence. The linearly polarized laser beam was converted into a circularly polarized beam with a quarter-wave plate. For photothermal therapy, cells were incubated with gold nanorods for a few hours and washed with fresh medium, and then fresh medium (0.3 mL) and ethidium bromide (EB) (2 mg mL1) in PBS (0.6 mL) were added into each well. The incident power was measured at the focus point. EB is a uorescence dye commonly used to stain nucleic acid and membrane impermeable for living cells. It can be used to examine the membrane integrity and viability of cells. In this work, EB stained on the nuclei of dead cells was imaged using two-photon excitation at a wavelength of 740 nm. The mean power density was calculated by dividing the average laser power with the scanning area (60 mm 60 mm). The energy uence was calculated by multiplying the mean power density (mean power/ scanning area) by the total exposure time t (exposure time per scan number of scans) of a nanorod. The exposure time of the nanorod was calculated following a known method. [13] For each scan with a duration of 1.05 s, the area covered was 60 mm 60 mm (512 pixels 512 pixels, pixel area 117 nm 117 nm). The exposure time per pixel was 4.0 ms. The focal spot area was calculated as pd2/4, where d 0.61 l/NA is the half width at half maximum of the beam waist. In one scan, the exposure time of an individual nanorod equal to (focal spot area/pixel area) 1.7 38.0 ms. The total irradiation time t 38.0 ms number of scans.

Received: April 7, 2008 Revised: May 23, 2008

[1] X. H. Huang, I. H. El-Sayed, W. Qian, M. A. El-Sayed, J. Am. Chem. Soc. 2006, 128, 2115. [2] T. B. Huff, L. Tong, Y. Zhao, M. N. Hansen, J. X. Cheng, A. Wei, Nanomedicine 2007, 2, 125. [3] A. M. Gobin, M. H. Lee, N. J. Halas, W. D. James, R. A. Drezek, J. L. West, Nano Lett. 2007, 7, 1929. [4] J. Y. Chen, D. L. Wang, J. F. Xi, L. Au, A. Siekkinen, A. Warsen, Z. Y. Li, H. Zhang, Y. N. Xia, X. D. Li, Nano Lett. 2007, 7, 1318. [5] H. F. Wang, T. B. Huff, D. A. Zweifel, W. He, P. S. Low, A. Wei, J. X. Cheng, Proc. Natl. Acad. Sci. USA 2005, 102, 15752. [6] T. K. Sau, C. J. Murphy, Langmuir 2004, 20, 6414. [7] P. Zijlstra, C. Bullen, J. W. M. Chon, M. Gu, J. Phys. Chem. B 2006, 110, 19315. [8] T. B. Huff, M. N. Hansen, Y. Zhao, J. X. Cheng, A. Wei, Langmuir 2007, 23, 1596. [9] N. J. Durr, T. Larson, D. K. Smith, B. A. Korgel, K. Sokolov, A. Ben-Yakar, Nano Lett. 2007, 7, 941. [10] D. Bird, M. Gu, Opt. Lett. 2003, 28, 1552. [11] L. Fu, A. Jain, C. Craneld, H. K. Xie, M. Gu, J. Biomed. Opt. 2007, 12. [12] J. Kim, S. Park, J. E. Lee, S. M. Jin, J. H. Lee, I. S. Lee, I. Yang, J. S. Kim, S. K. Kim, M. H. Cho, T. Hyeon, Angew. Chem. Int. Ed. 2006, 45, 7754. [13] L. Tong, Y. Zhao, T. B. Huff, M. N. Hansen, A. Wei, J. X. Cheng, Adv. Mater. 2007, 19, 3136. [14] H. Y. Li, Z. M. Qian, Med. Res. Rev. 2002, 22, 225. [15] Y. Y. Jiang, C. Liu, M. H. Hong, S. J. Zhu, Y. Y. Pei, Bioconjugate Chem. 2007, 18, 41.

3870

www.advmat.de

2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2008, 20, 38663871

COMMUNICATION

[16] P. H. Yang, X. Sun, J. F. Chiu, H. Sun, Q. Y. He, Bioconjugate Chem. 2005, 16, 494. [17] K. T. Yong, J. Qian, I. Roy, H. H. Lee, E. J. Bergey, K. M. Tramposch, S. He, M. T. Swihart, A. Maitra, P. N. Prasad, Nano Lett. 2007, 7, 761. [18] J. D. Bleil, M. S. Bretscher, EMBO J. 1982, 1, 351. [19] H. Ding, K. T. Yong, I. Roy, H. E. Pudavar, W. C. Law, E. J. Bergey, P. N. Prasad, J. Phys. Chem. C 2007, 111, 12552.

[20] C. P. Yao, R. Rahmanzadeh, E. Endl, Z. X. Zhang, J. Gerdes, G. Huttmann, J. Biomed. Opt. 2005, 10, 064012. [21] C. J. Engelbrecht, K. Greger, E. G. Reynaud, U. Krzic, J. Colombelli, E. H. K. Stelzer, Opt. Express 2007, 15, 6420. [22] American National Standard for Safe Use of Lasers ANSI Z136.1, American Laser Institute, Orlando, Florida, 2000. [23] C. Loo, A. Lowery, N. Halas, J. West, R. Drezek, Nano Lett. 2005, 5, 709.

Adv. Mater. 2008, 20, 38663871

2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advmat.de

3871