Cancer Photothermal Therapy with ICG-Conjugated Gold Nanoclusters

Abstract

The coming era of precision nanomedicine demands engineered nanoparticles that can be readily translated into the clinic, like that of molecular agents, without being hindered by intrinsic size heterogeneity and long-term body retention. Herein we report that conjugation of indocyanine green (ICG), an FDA-approved near-infrared (NIR) dye, onto an atomically precise glutathione-coated Au25 (GS-Au25) nanocluster led to a molecular-like photothermal nanoparticle (ICG₄–GS-Au25) with significantly enhanced ICG photostability and tumor targeting. Under weak NIR light irradiation conditions, free ICG failed to suppress tumor growth but the original tumors were completely eradicated with ICG₄–GS-Au25. In the meantime, “off-target” ICG₄–GS-Au25 was effectively cleared out from the body like small-molecule drugs after glutathione-mediated biotransformation in the liver. These findings highlight the merits of molecular-like nanomedicines, offering a new pathway to meet FDA’s criteria for the clinical translation of nanomedicines.

INTRODUCTION

Indocyanine green (ICG), currently the only FDA-approved near-infrared (NIR) dye, is widely used in many medical areas such as assessment of cardiovascular and liver function,¹ retinal angiography,² image-guided surgery³ as well as cancer phototherapy.^4,5 However, free ICG, with strong lipophilicity and serum protein affinity, suffers from rapid hepatobiliary elimination (plasma half-life <5 min) that severely limits its tumor targeting ability.⁶ To tackle this challenge, various nanoparticle delivery platforms were developed to improve the blood retention and tumor delivery efficiency of ICG.⁷ Among those delivery vectors, noble metal nanostructures have demonstrated great promise because of their various merits including bioinertness, ease of manufacturing, robustness, and tunable physiochemical properties that can further enhance the performance of ICG.^8,9 For instance, plasmonic gold nanoparticles,⁹ gold nanorods,¹⁰ and gold nanostars¹¹ as well as hollow gold nanospheres¹² all have been exploited as delivery platforms for ICG-based phototherapy. While these delivery systems generally exhibit an impressive photoinduced anticancer effect, their use in vivo and future clinical translation are significantly hampered by overwhelming and nonspecific accumulation in the mononuclear phagocyte system (MPS, e.g., liver and spleen). After MPS uptake, it has been shown that gold nanoparticles can be retained in the liver for over 6 months and induce adverse effects including oxidative stress, DNA damage, and inflammation, creating concerns over their long-term toxicity, which have been long-standing barriers for clinical translation of gold or other engineered nanoparticles.^13–15

Renal clearable gold nanoclusters is a special class of gold nanoclusters that highly resist serum protein adsorption and can be effectively eliminated through kidney glomerular filtration (size threshold ~6 nm) without massive MPS accumulation in vivo.^16–18 The unique physical and physiological properties of these ultrasmall nanoparticles have already led to many exciting biomedical applications that are otherwise unattainable by their larger counterparts.^19–21 Moreover, recent advancement in the chemical synthesis of renal clearable gold nanoparticles has enabled precise control of their sizes down to the single atom level^22–24, which not only ensures the batch-to-batch reproducibility that is crucial to their future clinical translation but also offers us unprecedented opportunities to probe the biological systems with the utmost precision.²⁵ It was also demonstrated recently that the ultrasmall size of renal clearable gold nanoparticles could allow for rapid penetration to the tumor cores and homogeneous distribution within the tumors, leading to an enhanced anticancer effect when compared to large nanoparticles as drug carriers.²⁶ Our previous results showed that conjugation of four ICG molecules onto an ultrasmall glutathione-coated Au25 nanocluster (ICG₄–GS-Au25) could effectively deliver ICG into tumor cells in vivo and hepatic glutathione-mediated biotransformation could be exploited to modulate the in vivo transport of engineered nanoparticles.²⁷

Herein we report that intravenously administered ICG₄–GS-Au25 could enable effective NIR light-induced tumor photothermal therapy, resulting in complete ablation of established tumors. Meanwhile, untargeted ICG₄–GS-Au25 nanoclusters were dissociated in the liver via biotransformation and the Au25 carriers were efficiently eliminated into the urine with minimal MPS accumulation (Scheme 1). Thus, integration of ICG with the renal clearable gold nanoparticle represents a promising strategy to enhance ICG-mediated cancer photothermal therapy while minimizing potential side effects associated with the nonspecific accumulation of delivery vectors.

Scheme 1.

ICG₄–GS-Au25-Mediated Cancer Photothermal Therapy and Its in Vivo Clearance Pathways after Dissociation in the Liver

RESULTS AND DISCUSSION

Characterization and Photothermal Properties of ICG₄–GS-Au25.

ICG₄–GS-Au25 was synthesized through a previously reported NHS/amine coupling reaction between NHS ester-activated ICG and the GS-Au25 nanoclusters (Figure 1A).²⁷ The products were further purified to yield GS-Au25 conjugated with an average of ~4 ICG molecules. TEM imaging confirms that the resulting ICG₄–GS-Au25 retains the ultrasmall core size of the Au25 nanocluster without aggregation in aqueous solution (Figure 1B), which is also verified by the dynamic light scattering (DLS) measurement (Supporting Information Figure S1). The absorption spectrum of ICG₄–GS-Au25 exhibits a dramatic blue-shift of the ICG absorption peak (from 795 to 710 nm) due to dipole–dipole interaction of multiple ICG dyes conjugated onto the same Au25 nanocluster (Figure 1C). Moreover, because of the extreme proximity between the Au25 and conjugated ICG dyes, the fluorescence of ICG is almost completely quenched after conjugation owing to the efficient photoinduced electron transfer (PET) process (Figure 1D). As the PET is strongly distance-dependent, the quenched ICG fluorescence can be instantaneously recovered once ICG is released from the surface of Au25 by other thiol molecules (Supporting Information Figure S2). In addition, only a minimal increase of ICG fluorescence was observed after incubation of ICG₄–GS-Au25 in PBS buffer or fetal bovine serum (FBS) for up to 8 h, suggesting the good stability of ICG₄–GS-Au25 in these physiological media (Supporting Information Figure S3).

Figure 1.

(A) Synthesis of ICG₄–GS-Au25 through simple NHS/amine coupling between ICG-NHS and GS-Au25. (B) TEM image of the synthesized ICG₄–GS-Au25 nanoclusters. (C) Absorption spectra of ICG₄–GS-Au25, free ICG, and GS-Au25 aqueous solutions. Inserted is a color picture of their aqueous solutions. (D) Fluorescence emission spectra of ICG₄–GS-Au25 and an equal amount of free ICG in aqueous solution excited at 760 nm. Insert is a scheme illustrating the ICG fluorescence quenching mechanism by Au25 through the photoinduced electron transfer (PET) process. (E) Fitted decay curves of the NIR absorption of ICG₄–GS-Au25 and equivalent free ICG aqueous solutions during irradiation with a 0.5 W/cm² 808 nm laser. (F) Normalized absorption spectra of ICG₄–GS-Au25 and free ICG aqueous solutions before and after 0.5 W/cm² 808 nm laser irradiation for 6 min. (G) Temperature change of ICG₄–GS-Au25 and an equivalent amount of free ICG or GS-Au25 in aqueous solution during 0.5 W/cm² 808 nm laser irradiation.

The photothermal properties of ICG₄–GS-Au25 was investigated with an 808 nm NIR diode laser. As shown in Figure 1E, under irradiation from a 0.5W/cm² NIR laser, the NIR absorption (700–900 nm area under the absorption spectrum) of free ICG decays rapidly with a half-life of only ~1.7 min while the NIR absorption of ICG₄–GS-Au25 decays much slower with a half-life (~6 min) over 3-fold longer than that of the free ICG. After 6 min laser irradiation, the NIR absorption of free ICG nearly disappears but ICG₄–GS-Au25 still exhibits relatively strong NIR absorption (Figure 1F and Supporting Information Figure S4). This enhanced photochemical stability of ICG conjugated to Au25 is likely due to the PET process (excited electron transfer from ICG to Au25) that inhibited the transfer of energy from the excited electron in ICG to nearby oxygen molecules, a process that is known to generate reactive singlet oxygen locally and expedite the photoinduced decomposition of ICG (Supporting Information Figure S5).²⁸ The photothermal effect of ICG₄–GS-Au25 was further evaluated in the test tube. Under 0.5W/cm² NIR laser irradiation, the temperature of the ICG₄–GS-Au25 solution (15 μM) increases rapidly by ~20 °C and maintains a high temperature with minimal decay during the 15 min irradiation (Figure 1G). In contrast, the equivalent ICG solution (60 μM) also shows a transient elevation of temperature (by ~10 °C) but decays rapidly afterward as a result of poor photochemical stability, while the temperature of the equivalent GS-Au25 solution only exhibits a slight increase because of its weak NIR absorption. These findings clearly indicated that the photothermal effect of ICG was enhanced after conjugating onto the GS-Au25 nanocluster.

Therapeutic Effect of ICG₄–GS-Au25 in Vitro.

We next studied the in vitro therapeutic effect of ICG₄–GS-Au25 in combination with an NIR laser using human breast cancer MCF-7 cells. As shown in Figure 2A, ICG fluorescence can be clearly observed inside the cells after 2 h incubation with ICG₄–GS-Au25, suggesting that ICG₄–GS-Au25 allows the efficient intracellular delivery of ICG. Additionally, the intracellular turn-on of ICG fluorescence indicates some ICG molecules can be released from Au25 after entering the cells, which is likely due to the high concentration of intracellular thiols (e.g., glutathione). Then the cells were incubated with 30 μM ICG₄–GS-Au25 and irradiated with the NIR laser at various power densities (0–1 W/cm²) for 5 min. The cell viability after irradiation was first directly visualized via live/dead cell staining. As shown in Figure 2B, cell death is laser power density-dependent and increases with the increase of power density until 0.8 W/cm², while the cells treated with PBS show negligible cell death regardless of the laser power density. This laser power density-dependent cytotoxicity of ICG₄–GS-Au25 was also confirmed by the MTT cell viability assay (Figure 2C). Under constant laser irradiation (0.5 W/cm²), cell death is also ICG₄–GS-Au25 concentration-dependent and a higher ICG₄–GS-Au25 incubation concentration results in greater cell death, which was visualized by the live/dead cell staining (Figure 2D) and quantified via the MTT assay (Figure 2E) as well. In the absence of laser irradiation, ICG₄–GS-Au25 did not show significant cytotoxicity in the range of concentration tested. Unexpectedly, we found that free ICG was over 6 times more cytotoxic than ICG₄–GS-Au25 as measured by IC₅₀ values under the same condition and laser treatment in vitro (Supporting Information Figure S6). This pronounced phototoxicity of free ICG most likely originated from the photodegradation products of free ICG because those products have been reported to be highly cytotoxic.²⁸ Nevertheless, combination of the results suggested that efficient cancer cell killing could be achieved through ICG₄–GS-Au25-mediated photothermal therapy.

Figure 2.

(A) Fluorescent microscope imaging of live MCF-7 cells at 2 h after incubation with ICG₄–GS-Au25. Scale bar, 10 μm. (B) Fluorescent microscope imaging of the viability of MCF-7 cells incubated with ICG₄–GS-Au25 (~30 μM) or PBS after irradiation with a 808 nm laser at different power densities. Live/dead cells were stained with green/red fluorescence by calcein-AM/propidium iodide. Scale bar, 100 μm. (C) Cell viabilities of experiment B quantified by MTT assay. (D) Fluorescent microscope imaging of the viability of MCF-7 cells incubated with various concentrations of ICG₄–GS-Au25 with or without 808 nm laser irradiation (0.5W/cm² for 5 min). Live/dead cells were stained with green/red fluorescence by calcein-AM/propidium iodide. Scale bar, 100 μm. (E) Cell viabilities of experiment D quantified by MTT assay.

ICG₄–GS-Au25-Mediated Photothermal Therapy in Vivo.

Because of the strong lipophilicity and protein affinity of ICG, ICG₄–GS-Au25 also strongly binds to serum proteins (Supporting Information Figure S7), which prevents rapid renal elimination of the ultrasmall ICG₄–GS-Au25. Conversely, the GS-Au25 nanocluster hinders the efficient liver uptake of the conjugated ICG, leading to over 12-fold enhancement in ICG blood retention time compared to that of free ICG (Figure 3A,B). This prolonged blood circulation is known to benefit passive tumor targeting through the enhanced permeability and retention (EPR) effect and allowed us to administer ICG₄–GS-Au25 intravenously for efficient tumor targeting (Supporting Information Figure S8).²⁷ At 5 h after iv injection of ICG₄–GS-Au25 to BALB/c mice bearing breast cancer, the photothermal treatment of the tumor was conducted using the 808 nm diode laser at a power density of 0.8 W/cm² for 8 min. As shown in Figure 3C,D, under laser irradiation the tumor temperature of ICG₄–GS-Au25 injected mice surged by over 20 °C within 3 min to ~55 °C and maintained a high temperature for the next 5 min. In comparison, the tumor temperature of mice injected with free ICG only increased by less than 10 °C, which was similar to that of the PBS-injected group. This poor photothermal effect of free ICG in vivo is fundamentally due to its rapid uptake by the liver, which severely limits its blood retention and tumor targeting. Noteworthily, the laser-induced tumor hyperthermia in mice injected with ICG₄–GS-Au25 is comparable to that achieved with long-circulating PEGylated Au nanostructures of strong plasmonic NIR absorption (peak ~800 nm) at similar doses.²⁹ After photothermal treatment, the tumor volumes of ICG₄–GS-Au25-injected mice continued to shrink and completely disappeared at ~2 week post treatment (Figure 3E,F); however, the free ICG-injected mice failed to show any therapeutic efficacy and the tumor growth kinetics was nearly identical to that of the PBS-injected group, consistent with their inferior tumor hyperthermia effect. These results demonstrated that ICG₄–GS-Au25 could serve as an excellent intravenous photothermal agent for effective cancer photothermal therapy.

Figure 3.

(A) 24 h ICG blood pharmacokinetics in mice iv injected with either ICG₄–GS-Au25 or an equivalent amount of free ICG. (B) Area under the pharmacokinetics curves (AUCs) of ICG₄–GS-Au25 and free ICG. (C) Tumor temperature kinetics of mice intravenously injected with ICG₄–GS-Au25, free ICG, or PBS during 0.8W/cm² 808 nm laser irradiation for 8 min. n = 3 mice for each group. (D) Representative tumor thermal images of the mice receiving PTT treatment at 8 min of laser irradiation. White arrows indicate the tumors on mice. (E) Representative color images of the tumors on mice at different time points post PTT treatment. (F) Tumor volume kinetics of mice after PTT treatment. Statistical analysis was performed using two-sample equal-variance t test (**P < 0.005, ***P < 0.0005).

Biocompatibility and Toxicity Evaluation.

Despite being outstanding photothermal transducers and extensively investigated, plasmonic noble metal nanostructures are still far from clinical in vivo application, as they cannot efficiently clear from the body and often accumulate overwhelmingly in MPS organs such as the liver and spleen for prolonged time.^15,29 Though ICG₄–GS-Au25 nanoclusters bind to serum proteins, they were found to escape MPS uptake and clear out to urine with over 50% ID at 24h pi after dissociation from the protein-binding ICG in vivo (Figure 4A and Supporting Information Figure S9). Immediately after intravenous injection of ICG₄–GS-Au25, activation of ICG fluorescence can be observed from the liver while the fluorescence activation in extrahepatic circulation is minimal, indicating that ICG₄–GS-Au25 is stable in extrahepatic circulation but undergoes hepatic glutathione-mediated dissociation inside the liver (Supporting Information Figure S10), which is the primary organ responsible for biotransformation of xenobiotics. Because of this in vivo dissociation and efficient renal clearance of Au25, the Au accumulation in major organs were minimized and continued to decrease over time after the injection of ICG₄–GS-Au25 (Supporting Information Figure S11), in sharp contrast to the reported large plasmonic photothermal nanostructures that accumulate predominantly and persistently in the liver and spleen following intravenous injection.²⁹ Meanwhile, dissociated ICG was still eliminated through its inherent hepatobiliary pathway to feces (Supporting Information Figure S12).

Figure 4.

(A) Renal clearance efficiency of intravenously injected ICG₄–GS-Au25 in mice. Inserts are color pictures of the injected ICG₄–GS-Au25 solution, excreted urine exhibiting the typical color of Au25 nanoclusters as well as GS-Au25 PBS solution and control mouse urine for reference. (B) Body-weight trend of mice (n = 3) injected with ICG₄–GS-Au25, free ICG, or PBS after PTT treatment. (C) Serum alanine aminotransferase (ALT) levels of ICG₄–GS-Au25 injected mice at 24 h and 7 day pi. (D) Serum albumin levels of ICG₄–GS-Au25 injected mice at 24 h and 7 day pi. (E) Blood urea nitrogen (BUN) levels of ICG₄–GS-Au25 injected mice at 24 h and 7day pi. (F) Serum creatinine levels of ICG₄–GS-Au25 injected mice at 24 h and 7 day pi. (G) BUN/creatinine ratios of ICG₄–GS-Au25 injected mice at 24 h and 7 day pi. (H) Histopathology of major organs of ICG₄–GS-Au25 injected mice at 24 h and 7day pi. Scale bar, 150 μm. Statistical analysis was performed using two-sample equal-variance t test (n.s., nonsignificant difference, P > 0.05).

The mice were monitored periodically after receiving photothermal treatment and did not show significant behavior change or loss of body weight in the 2 weeks following treatment (Figure 4B), suggesting good biocompatibility of the ICG₄–GS-Au25 as a photothermal agent. Because the liver and kidneys are the main organs for biotransformation and elimination of ICG₄–GS-Au25 in vivo, we further tested the blood chemistry to assess the hepatotoxicity and nephrotoxicity. Serum alanine aminotransferase (ALT) (Figure 4C) and albumin (Figure 4D) levels at 24 h and 7 d after ICG₄–GS-Au25 administration were statistically the same as control values, which indicated that ICG₄–GS-Au25 did not induce obvious hepatotoxicity. No apparent nephrotoxicity was found either, as the blood urea nitrogen (BUN) (Figure 4E), creatinine (Figure 4F), and BUN-to-creatinine ratio (Figure 4G) at 24 h and 7 d after ICG₄–GS-Au25 administration were not significantly different from those of the control. Furthermore, histological and pathological analysis of major organs at 24 h and 7 d postinjection of ICG₄–GS-Au25 did not show any abnormality (Figure 4H). Collectively, this evidence supports that ICG₄–GS-Au25 is highly biocompatible and safe when used as an intravenous agent for cancer photothermal therapy.

CONCLUSION

In summary, we have demonstrated that the ultrasmall renal-clearable atomically precise GS-Au25 nanocluster is a promising vector to deliver ICG, the FDA-approved NIR dye, for more effective cancer photothermal therapy while simultaneously avoiding potential side effects associated with the long-term accumulation of carriers in healthy tissues. The resulting ICG₄–GS-Au25 could increase the photochemical stability of ICG, improve its photothermal performance, prolong its blood circulation, and enhance its tumor targeting, which all contributed to the significantly improved photothermal therapeutic efficacy. Moreover, “off-target” ICG₄–GS-Au25 could take advantage of hepatic biotransformation to dissociate inside the liver; as a result, the detached Au25 nanoclusters were cleared out efficiently through the urine while the detached ICG was still eliminated through its inherent hepatobiliary pathway, resulting in the minimized accumulation of carriers in normal tissues and excellent biocompatibility. Despite extensive investigation of plasmonic nanostructures as photothermal agents in the past decades, few of them can translate to the clinic due to difficulties in reproducible synthesis caused by size heterogeneity and the potential toxicity associated with the long-term accumulation in healthy organs. Our findings of the molecular-like dissociable and clearable ICG₄–GS-Au25 highlighted synergistic effects arising from integration of molecular agents with atomically precise nanoclusters, which offer a new pathway to meet FDA’s criteria and overcome the barriers in the clinical translation of nanomedicines.

MATERIALS AND METHODS

Synthesis of ICG₄–GS-Au25.

Atomically precise glutathione-coated Au25 nanoclusters (GS-Au25) was synthesized based on reported method.³⁰ The synthesis of ICG₄–GS-Au25 was also reported in our previous publication.²⁷ Briefly, 4 mg of ICG-NHS ester (Intrace medical, Switzerland) dissolved in 6 mL of DMSO was added to 4 mL of aqueous solution containing 6 mg of GS-Au25 nanoclusters. The mixture was then vortexed for 3 h at room temperature, and the resulting conjugates were precipitated by centrifugation with the copious addition of ethanol. The conjugates were again redispersed in 1X PBS buffer and purified with 30KDa Amicon Ultra centrifuge filters (Sigma) to remove any unconjugated GS-Au25 nanoclusters. The resulting ICG₄–GS-Au25 inside the centrifuge filter was resuspended in ultrapure water and lyophilized for future usage.

Evaluation of ICG₄–GS-Au25 Photothermal Effect in Test Tubes.

ICG₄–GS-Au25 (10 μM) was suspended as an aqueous solution and added to a 5 mm width quartz cuvette, which was exposed to 0.5W/cm² 808 nm laser irradiation. The laser power density was measured with an optical power meter kit (Thorlabs PM100). The solution temperature was remotely recorded with an infrared thermal camera (FLIR E6). Free ICG (40 μM) and GS-Au25 (10 μM) solutions were measured at the same conditions for comparison. For the photochemical stability test, ICG₄–GS-Au25 (1 μM) and free ICG (4 μM) solutions were placed in a 1 mm width quartz cuvette and received 0.5W/cm² 808 nm laser irradiation, followed by measurement of their absorption spectra with a UV/vis spectrometer (Varian Cary 50 Bio) at specific time points.

Evaluation of ICG₄–GS-Au25 Photothermal Therapy Effect in Vitro.

Human breast cancer MCF-7 cells were seeded in 96-well plates and incubated until the cells reached ~70–80% confluency. Then the cell growth medium in the wells was replaced by different concentrations of ICG₄–GS-Au25 dissolved in MEM medium, and the cells were further incubated for 2 h at 37 °C in the cell incubator, followed by 808 nm laser irradiation at various power densities. The cell viability in each vial after irradiation was directly visualized under a fluorescence microscope (Olympus IX71 with optical filter sets and CCD camera) using calcein AM/propidium iodide (PI) for live/dead cell staining. The cell viability in each well was also confirmed by the standard 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay using a microplate reader (Molecular Devices EMax Plus).

Evaluation of ICG₄–GS-Au25 Photothermal Therapy Effect in Vivo.

Female BALB/c mice (6–8 weeks old, body weight ~25g) were inoculated subcutaneously with ~10⁵ mouse breast cancer cells (TUBO)/mouse, and the tumors were allowed to grow ~2 weeks until they reached a volume of ~100 mm³ (width² × length/2). Then the tumor-bearing mice (n = 3) were intravenously injected with 150 μL 1.5 mM ICG₄–GS-Au25 PBS solution/mouse. The control groups received an equivalent dose of free ICG (n = 3 mice) and PBS (n = 3 mice) solution. At 5 h after injection, mice were anesthetized with 3% isoflurane and placed on an adjustable stage with the tumor facing upward. The tumor areas were then continuously irradiated with a 808 nm laser at a power density of 0.8 W/cm² for 8 min. Meanwhile, the tumor temperature was recorded with an infrared thermal camera at multiple time points during the laser treatment. After this single treatment, the tumor volume of each mouse was measured every 2 days with a caliper for 16 consecutive days.

Biodistribution of Au and in Vivo Toxicity Evaluation.

BALB/c mice (n = 3 mice at each time point) were sacrificed at specific time points after intravenous injection of 150 μL of 1.5 mM ICG₄–GS-Au25 PBS solution/mouse, and the organs were immediately collected and immersed in 10% neutral buffered formalin for over 48 h. Then the tissues were processed by standard dehydration and paraffin embedding. The embedded tissues were further sectioned into 4 μm slices and H&E stained for pathological assessment. Blood was withdrawn via cardiac puncture during the sacrifice process, and the obtained serum samples were stored in −20 °C until they were ready for biomarker tests. For biodistribution study, the collected organs were completely digested in freshly prepared aqua regia for 72 h. Then the Au concentration in the organs was accurately quantified by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7900) after proper dilution.