Gold-Nanoclustered Hyaluronan NanoAssemblies for Photothermally Maneuvered Photodynamic Tumor Ablation Hwa Seung Han,†,§ Ki Young Choi,†,§ Hansang Lee,† Minchang Lee,† Jae Yoon An,† Sol Shin,† Seunglee Kwon,† Doo Sung Lee,† and Jae Hyung Park*,†,‡ †
School of Chemical Engineering, Sungkyunkwan University, Suwon 16419, Republic of Korea Department of Health Sciences and Technology, SAIHST, Sungkyunkwan University, Seoul 06351, Republic of Korea
‡
S Supporting Information *
ABSTRACT: Optically active nanomaterials have shown great promise as a nanomedicine platform for photothermal or photodynamic cancer therapies. Herein, we report a gold-nanoclustered hyaluronan nanoassembly (GNc-HyNA) for photothermally boosted photodynamic tumor ablation. Unlike other supramolecular gold constructs based on gold nanoparticle building blocks, this system utilizes the nanoassembly of amphiphilic hyaluronan conjugates as a drug carrier for a hydrophobic photodynamic therapy agent verteporfin, a polymeric reducing agent, and an organic nanoscaffold upon which gold can grow. Gold nanoclusters were selectively installed on the outer shell of the hyaluronan nanoassembly, forming a gold shell. Given the dual protection effect by the hyaluronan self-assembly as well as by the inorganic gold shell, verteporfin-encapsulated GNc-HyNA (Vp-GNc-HyNA) exhibited outstanding stability in the bloodstream. Interestingly, the fluorescence and photodynamic properties of Vp-GNcHyNA were considerably quenched due to the gold nanoclusters covering the surface of the nanoassemblies; however, photothermal activation by 808 nm laser irradiation induced a significant increase in temperature, which empowered the PDT effect of Vp-GNc-HyNA. Furthermore, fluorescence and photodynamic effects were recovered far more rapidly in cancer cells due to certain intracellular enzymes, particularly hyaluronidases and glutathione. Vp-GNc-HyNA exerted a great potential to treat tumors both in vitro and in vivo. Tumors were completely ablated with a 100% survival rate and complete skin regeneration over the 50 days following Vp-GNc-HyNA treatment in an orthotopic breast tumor model. Our results suggest that photothermally boosted photodynamic therapy using Vp-GNc-HyNA can offer a potent therapeutic means to eradicate tumors. KEYWORDS: hybrid nanomaterials, gold, hyaluronan, photothermal therapy, photodynamic therapy
O
antibodies, or therapeutic/imaging agent to provide the gold nanomaterial with functionalities like disease-specific targetability or therapeutic activity.5−7 Recent studies demonstrated an advanced approach to the preparation of gold-based hybrid nanomaterials, where relatively small gold nanoparticles (530 nm) and amphiphilic polymers were employed as building blocks to prepare micelle/ vesicle-type supramolecular self-assemblies.8−11 These studies employed gold nanoparticles as a scaffold or a basic building block for constructing supramolecular structures of gold nanoparticles, wherein the organic compartments generally serve as a connecting material, enabling gold nanoparticle selfassembly. Since spherical gold nanoparticles generally have an
rganic/inorganic hybrid nanomaterials have been intensively investigated in the biomedical research field because they benefit not only from unique attributes of inorganic compartments like physical and chemical stability, bioinertness, and interesting optical properties1,2 but also from those of organic compartments such as biodegradability, plasticity, and flexibility of molecular design.3,4 Among the diverse hybrid nanomaterials, gold-based hybrid nanomaterials have shown great promise as a nanomedicine platform, because gold nanomaterials are considered nontoxic and biocompatible; they also have exciting optical properties such as a surface plasmon resonance (SPR) effect. To synthesize gold-based hybrid nanomaterials, gold nanomaterials such as gold nanoparticles, nanorods, or nanostars with diameters in the range of 50−200 nm have been widely used as core materials, and their surfaces have been modified with functional molecules like targeting moieties, for example, peptides, © 2016 American Chemical Society
Received: July 30, 2016 Accepted: October 24, 2016 Published: November 16, 2016 10858
DOI: 10.1021/acsnano.6b05113 ACS Nano 2016, 10, 10858−10868
Article
www.acsnano.org
Article
ACS Nano
Figure 1. Synthesis and physicochemical characterization of gold-nanoclustered hyaluronan nanoassemblies (GNc-HyNA). (A) Schematic illustration for synthetic steps and photothermal effects of GNc-HyNA upon 808 nm PTT laser irradiation. (B) TEM images of HyNA and GNc-HyNA fabricated with 0.2, 0.5, and 0.9 μM concentrations of Au-seed. Scale bar = 100 nm. (C) Photographs of GNc-HyNA at various concentrations of Au-seed. (D) UV−vis-NIR spectra of HyNA and GNc-HyNA with 0.2, 0.5, and 0.9 μM of Au-seed. (E) TEM images of GNcHyNA after PTT laser irradiation. Scale bar = 100 nm. (F) Photothermal images of GNc-HyNA and HyNA under PTT laser irradiation. Thermal heating curves of (G) GNc-HyNA with different Au-seed concentrations under the PTT laser irradiation or (H) GNc-HyNA (0.9 μM) with different concentrations (0.1, 0.5, and 1 mg/mL) under 808 nm PTT or 690 nm PDT laser irradiation.
Figure 2. Photoactivatable characteristics of verteporfin (Vp)-loaded GNc-HyNA (Vp-GNc-HyNA). (A) Schematic illustration of the Vp release mechanism from Vp-GNc-HyNA by sequential PTT/PDT laser treatment. (B) Time-dependent change in relative scattering intensity of Vp-GNc-HyNA and Vp-HyNA in a 50% FBS solution (1 mg/mL). (C) Fluorescence intensity of Vp, Vp-HyNA, and Vp-GNc-HyNA (1 mg/ mL) before and after 808 nm PTT laser irradiation. (D) In vitro release profiles of Vp fromVp-HyNA and Vp-GNc-HyNA (1 mg/mL) in the absence or presence of PTT laser irradiation. (E) Singlet oxygen generation from Vp-HyNA and Vp-GNc-HyNA (100 μg/mL of Vp) in the presence of PTT laser and/or PDT laser irradiation. The error bars in the graph indicate standard deviations (n = 5). Asterisks (*) denote statistically significant differences (*p < 0.05) calculated via one-way ANOVA.
10859
DOI: 10.1021/acsnano.6b05113 ACS Nano 2016, 10, 10858−10868
Article
ACS Nano
RESULTS AND DISCUSSION For preparation of the HyNA, we synthesized an amphiphilic block copolymer by conjugating azide-functionalized PCL to αalkyne Hy conjugated with dimethylaminopropylamine (DAA)using facile click chemistry. In detail, an alkyne group was introduced to Hy by reductive amination with propargylamine in the presence of the reducing agent sodium cyanoborohydride (Supporting Information Figure S1). The 1H NMR spectrum confirmed the presence of both Hy and PCL, which indicates successful formation of block copolymers (Supporting Information Figure S2). DAA, which was chosen as a reducing moiety for formation of gold, was conjugated to the backbone of Hy through amide-bond formation. The successful synthesis of DAA-conjugated Hy-b-PCL was confirmed based on characteristic peaks of Hy (2.0 ppm), PCL (1.0 ppm), and DAA (2.2 ppm). The degree of substitution, defined as the number of DAA derivatives per 100 Hy repeating units, was found to be 31%. Owing to its amphiphilicity, the DAAmodified Hy-PCL conjugate could readily form micellar selfassemblies under aqueous conditions. Hy-PCL conjugates were self-assembled above the critical micelle concentration (CMC) of 72 μg/mL, which is significantly lower than that of low molecular weight detergent (Supporting Information Figure S3).23 After preparing HyNA, we deposited gold nanoclusters on the HyNA surface by reducing gold-seed (Au-seed) using the polymeric reducing agent DAA-HyNA. To prepare GNcHyNA, we added HAuCl4 to HyNAs, on which gold ions nucleated and grew to form gold nanoclusters on the surface of the HyNAs; the gold nanoclusters were stabilized with the tertiary amine chains of the DAA groups.24−26 Successful anchoring of gold nanoclusters was confirmed by UV−vis spectroscopy as well as by energy dispersive X-ray photoelectron spectroscopy, which showed a pronounced peak for gold (Figure 1C,D and Supporting Information Figure S4B). The formation of gold nanoclusters on HyNAs enabled us to control the physicochemical properties of GNc-HyNAs. TEM images showed that both HyNAs and GNc-HyNAs were spherically shaped, but GNc-HyNA seemed much denser and to have uneven edges and a more compact shape compared to uncoated HyNAs (Figure 1B). The hydrodynamic diameters of HyNA and GNc-HyNA were found to be 204.0 ± 5.4 and 150.2 ± 3.1 nm, respectively (Supporting Information Table S1). The decrease in a diameter after installation of gold nanoclusters on HyNAs could be attributed to physical crosslinking with the gold nanoclusters, thereby forming a compact nanostructure. In addition to the physicochemical characteristics, the optical properties of GNc-HyNA were also tailored with various amounts of HAuCl4 added to DAA-decorated HyNA. When 0.5 μM or a lower amount of Au-seed was added to the HyNA, a plasmon resonance peak around 550 nm was observed. However, when more than 0.9 μM of Au-seed was added, a new broad peak centered at 780 nm with ranges between 650 and 900 nm appeared due to the aggregation of Au nanoparticles.9,27,28 Thus, the photothermal effect of GNcHyNA was investigated. GNc-HyNA solutions were exposed to PTT laser at a power density of 1.5 W/cm2, which is known as a moderate power in PTT applications.9,29,30 GNc-HyNA added with 0.2 μM or 0.5 μM Au-seed showed a poor photothermal effect, and the temperature of the solutions slightly increased up to 39.1 or 33.9 °C, respectively, because
SPR peak in the range of 500−550 nm, their in vivo application to photothermal therapy is limited due to the strong absorption by tissue, particularly hemoglobin, in the visible region.12 However, with gold nanoparticles agglomerating into supramolecular assemblies, the SPR peak can red-shift to the nearinfrared (NIR) region; hemoglobin and water have their lowest absorption coefficient in the NIR region, which can facilitate photothermal therapy (PTT) in vivo. Recently, a variety of gold-based self-assemblies have been proposed as therapeutic platforms for drug delivery,9,13 PTT,8,9 and PDT/PTT dual therapy.8,14−16 Herein, we report a biostable, photoactivatable hybrid nanomaterial based on a gold-nanoclustered hyaluronan nanoassembly (GNc-HyNA) to achieve complete tumor eradication empowered by sequential therapeutics of plasmonic PTT and photothermally boosted PDT. For the preparation of the hybrid nanomaterials in this study, we employed the polymeric nanoassembly of amphiphilic hyaluronan-polycaprolactone (Hy-PCL) conjugates as a polymeric reducing agent instead of using small molecule reducing agents such as sodium citrate or sodium borohydride. Unlike other supramolecular gold constructs based on metallic gold nanoparticle building blocks, this system uses nanoassembly as a polymeric nanoscaffold, which the gold nanoclusters can grow on and hydrophobic verteporfin (Vp), an FDA approved PDT agent,17,18 can abide in (Figure 1A). To grow gold nanoclusters on the polymeric self-assemblies, we did not exploit any surfactant such as cetyltrimethylammonium bromide. Since the nanoassembly is comprised of the hydrophilic hyaluronan (Hy) corona and hydrophobic PCL core, hydrophobic Vp molecules were able to be encapsulated in the hydrophobic PCL core. The Hy corona of the polymeric scaffold can also act as an active targeting ligand to facilitate CD44 mediated tumortargeting and cell permeation because the Hy receptor CD44 is known to be overexpressed on various tumor cells.19,20 Given the tertiary amine moieties introduced onto the Hy outer layer of the nanoassembly, gold nanoclusters can be selectively installed on the surface of the nanoassembly forming a gold shell. The gold shell acts as a diffusion barrier to impede premature release of Vp molecules from Vp-encapsulated GNcHyNA (Vp-GNc-HyNA) in the bloodstream, and it also serves as an optical barrier to protect the photoreactive Vp from external light, hindering singlet oxygen generation. It should be noted that hampering untimely release of Vp and singlet oxygen generation under physiological conditions reduces potential toxicity of the PDT agent to superficial tissues such as the skin or eyes upon exposure to sunlight after administration. PTT laser irradiation following GNc-HyNA treatment can induce significant increases in local temperature by a photothermal effect; the soaring heat can lead to dissociation of the Vp-GNc-HyNA, which markedly accelerates Vp release from Vp-GNc-HyNA and thus promptly activates the photodynamic effect upon PDT laser radiation. The photothermally boosted Vp release is even faster in cancer cells because Hy molecules can be degraded in the cells by hyaluronidases (Figure 2A); Hy enzymes known to be overexpressed in many types of cancer cells.21,22 In this report, we demonstrated (i) selective deposition of gold nanoclusters on the outer shell of a drug-loaded hyaluronan nanoassembly (HyNA) and (ii) the application of gold-nanoclustered polymeric nanoassemblies for photothermally activated photodynamic therapy. 10860
DOI: 10.1021/acsnano.6b05113 ACS Nano 2016, 10, 10858−10868
Article
ACS Nano
Figure 3. In vitro cellular uptake and release profiles with treatment of enzymes and laser. (A) Confocal microscopy images of MDA-MB-231 cells incubated with Cy5.5-labeled GNc-HyNA (100 μg/mL) for 3 h. (B) Intracellular internalization of Vp-GNc-HyNA (30 μg/mL) into MDA-MB-231 cells after a 3 h post-treatment period. (C) In vitro Vp release behavior from Vp-GNc-HyNA (1 mg/mL) in the presence or absence of enzymes and the PTT laser treatment. (D) Singlet oxygen generation from Vp-Gc-HyNA (100 μg/mL of Vp) in the presence and absence of enzymes and/or laser treatments. (E) Cell viability of MDA-MB-231 cells incubated with different concentrations of Vp-GNcHyNA for 6 h in the presence of laser treatments. The error bars in the graph indicate standard deviations (n = 5). Asterisks (*) denote statistically significant differences (*p < 0.05) calculated via one-way ANOVA.
their SPR peaks were found in the range of 450 to 600 nm. Significant absorbance was not detected in the NIR region. By contrast, the temperature of the GNc-HyNA 0.9 μM solution soared to 74.1 °C given its strong SPR peak in the NIR region (Figure 1F−H). As the concentration of GNc-HyNA increased from 0.1 to 1 mg/mL, the solution temperature rose from 44 to 74 °C after PTT laser irradiation for 10 min, whereas temperature change for pure water exposed to PTT laser or Vp-GNc-HyNA solutions under PDT laser irradiation was negligible. The broad and red-shifted SPR spectra and the remarkable temperature increases indicated that GNc-HyNA with 0.9 μM of Au-seed would be suitable for NIR-based photothermal applications of GNc-HyNAs. In this study, we aimed to prepare a potent nanomedicine platform that can be employed as a robust drug delivery carrier as well as a photothermally activatable PDT agent. To develop the nanomedicine platform, GNc-HyNA was exploited as a drug delivery carrier because the PCL-based hydrophobic inner core of HyNA can act as a drug reservoir. The PDT agent, Vp, was encapsulated in the PCL core, and gold was sequentially deposited on the surface of Vp-HyNA (Figure 2A). When 10 wt % of Vp was added, 90% and 87% of the added Vp was encapsulated into Vp-HyNA and Vp-GNc-HyNA, respectively (Supporting Information Table S2). To investigate the structural stability of Vp-GNc-HyNA, we monitored the changes in scattering intensities of the Vp-GNc-HyNA and
Vp-HyNA under serum-mimicking conditions (50% FBScontaining PBS) (Figure 2B). The Vp-GNc-HyNA maintained 63% of the initial intensity up to 9 h. However, the scattering intensity of Vp-HyNA steeply decreased by 26% of the initial intensity. These results suggest the Vp-GNc-HyNA can effectively maintain their structural integrity in a physiological environment and, thus, can securely protect drug molecules, which is one of the most important requisites for drug delivery carriers to efficiently deliver drugs to their target sites under physiological conditions. To investigate the photophysical and photochemical properties of Vp-GNc-HyNA, the fluorescence intensity of Vp-GNcHyNA was observed using a fluorescence spectrophotometer. Notably, when Vp was loaded in GNc-HyNA, its intrinsic fluorescence signals were highly quenched due to the strong surface plasmonic effect of the gold nanoclusters (Figure 2C).31,32 The pronounced quenching effect of Vp-GNc-HyNA was not confined to the fluorescence signals, but singlet oxygen generation was also conspicuously diminished in the GNcHyNA. To explore the ability of Vp-GNc-HyNA to generate singlet oxygen via the photodynamic effect, the nanoassemblies in an aqueous solution were irradiated with a 690 nm PDT laser, and singlet oxygen generation was monitored using pnitrosodimethylaniline (RNO) as a singlet oxygen sensor.33 The RNO concentration was recorded as a function of time under various conditions (Figure 2 E). Singlet oxygen 10861
DOI: 10.1021/acsnano.6b05113 ACS Nano 2016, 10, 10858−10868
Article
ACS Nano
Figure 4. In vivo NIRF imaging of GNc-HyNA. (A, D) Time-lapse fluorescence imaging of GNc-HyNA (1 mg/mL) and ex vivo organ imaging 24 h after injection into subcutaneous MDA-MB-231 tumor models (A) and orthotopic tumor models (D). (B, E) Fluorescence intensity of Cy5.5-GNc-HyNA in the whole body and the tumor site of subcutaneous tumor models (B) or orthotopic tumor models (E) as a function of time. (C, F) Quantification of fluorescence intensity in tumor and major organs of subcutaneous tumor models (C) or orthotopic tumor models (F). (G) Thermal images of MDA-MB-231 tumor-bearing mice 9 h postinjection of PBS and GNc-HyNA (1 mg/mL) under the PTT laser irradiation. (H) Ex vivo organ distribution of Au after intravenous injection of GNc-HyNA in to subcutaneous tumor models. (I) Thermal heating curve of GNc-HyNA (1 mg/mL) on tumors as a function of PTT laser treatment time.
generation from Vp was impeded by a self-quenching effect when Vp was tightly aggregated in the micelle core as previously reported.34,35 In addition, gold nanoclusters on the GNc-HyNA hindered singlet oxygen generation of Vp due to energy transfer from the excited Vp to gold nanoclusters (Figure 2E).16 When Vp molecules were encapsulated by GNcHyNA, PDT laser radiation could not effectively generate singlet oxygen. This interesting PDT quenching phenomenon of Vp-GNc-HyNA not only indicates secure Vp loading in the nanoassemblies but also implies a possibility of minimizing nonspecific toxicity of PDT to patients who received PDT treatment upon exposure to sunlight. For Vp-GNc-HyNA to be an effective nanomedicine platform, it must not generate toxic reactive oxygen species in the bloodstream or normal organs; however, it needs to effectively produce singlet oxygen at disease sites like tumor
cells. Therefore, we demonstrated a further therapeutic application of Vp-GNc-HyNA that enables activatable, on-site photodynamic therapeutics. First, Vp-release profiles were monitored with Vp-GNc-HyNA incubated in PBS at 37 °C. The release rate of Vp from Vp-GNc-HyNA was slowed under physiological conditions (Figure 2D); only 30.8% of Vp was released in PBS within 48 h. However, the Vp release from VpGNc-HyNA under PTT laser irradiation dramatically accelerated. More than 73% of encapsulated Vp was released under exposure to the PTT laser at 48 h. Importantly, sequential radiation of PTT/PDT laser onto Vp-GNc-HyNA greatly boosted singlet oxygen generation as compared with that under single PDT laser treatment. This suggests that the gold shell of the Vp-GNc-HyNA significantly quenched the fluorescence signals of and suppressed singlet oxygen generation from Vp. Under sequential exposure to PTT/PDT laser, however, there 10862
DOI: 10.1021/acsnano.6b05113 ACS Nano 2016, 10, 10858−10868
Article
ACS Nano
Vp-GNc-HyNA on MDA-MB-231 cells (Figure 3E). As expected, the GNc-HyNA or Vp-GNc-HyNA did not induce noticeable cytotoxicity without laser irradiation (Supporting Information Figure S5A,B). However, substantial cytotoxicity was induced in cells exposed to PDT, PTT, or sequential PDT/ PTT laser treatment. Interestingly, Vp-GNc-HyNA treated sequentially with PTT laser and PDT laser irradiation led to even stronger cytotoxicity than Vp-HyNA under PDT laser, Vp-GNc-HyNA under PTT, or PDT laser alone (Figure 3E). Treatment with Vp-GNc-HyNA and sequential PTT/PDT laser radiation resulted in the highest anticancer efficacy among the treatment groups. We evaluated the in vivo behavior of Vp-GNc-HyNA. For the in vivo study, we established subcutaneous tumor models and orthotopic breast tumor models prepared by subcutaneously or orthotopically injecting MDA-MB-231 human breast cancer cells into the mammary gland of Balb/c nude mice; these models are considered more clinically relevant and better predictive models for drug evaluation than subcutaneous models. First, we monitored time-dependent biodistribution of the nanoparticles using in vivo NIR fluorescence microscopy after intravenous administration of GNc-HyNA labeled with Cy5.5 into both subcutaneous and orthotopic breast tumor models (Figure 4A,D). Fluorescence signals from the GNcHyNA were found all over the body 1 h postinjection, and the signals gradually diminished as time passed; thus, only feeble fluorescence signals were found in the body 24 h postinjection. Interestingly, stronger fluorescence signals were detected at tumor sites 1 h after injection than seen in the whole body. The fluorescence signals gradually increased and reached the highest level at 9 h (Figure 4B,E). It is important to note that the strong fluorescence intensities in the tumor remained steady at 24 h, whereas those from the whole body rapidly decreased. These results imply that GNc-HyNA can circulate in the bloodstream and rapidly target tumors. While GNc-HyNA is readily excreted from the body, it can stay in tumor tissues for a prolonged period of time, possibly due to impaired lymphatic drainage in tumor tissues.43,44 To evaluate the relative organ and tumor distribution of GNc-HyNA, the major organs of subcutaneous tumor xenograft mice and orthotopic breast tumor mice were harvested 2 days after injection, and the fluorescence signals of the organs under an in vivo imaging system were examined (Figure 4C,F). Ex vivo organ images showed that GNc-HyNA could accumulate at high levels in tumor, liver, and kidney tissues, but the tumor accumulation was greater than that of the other organs. The gold accumulated in the major organs was quantified by ICPMS at different time points (Figure 4H). We found that GNcHyNA accumulated prominently in tumor, liver, and spleen tissues on day 1. More than 25.1% of injected gold was found in tumor tissues, and 36.1% and 16.9% were detected in liver and spleen tissues, respectively. Compared to the gold contents on day 1, accumulated gold contents rapidly decreased in the organs as the day progressed; over 69.8% and 50.2% of the gold content accumulated in liver and spleen tissues were cleared by day 7, and over 86.1% and 92.9% were cleared by day 14. On the other hand, over 56.1% of gold accumulated in tumor tissues remained by day 7, and 20.4% remained on day 14, which demonstrated prolonged retention effect of nanostructures in tumor tissues. We further confirmed tissue distribution of the PDT agent Vp delivered with GNc-HyNA by detecting fluorescence signals of Vp at both tumors and organs (Supporting Information Figure S7).
was rapid Vp release followed by a remarkable recovery of both fluorescence signals and singlet oxygen generation, which consequentially augmented the PDT efficacy. One of the biggest advantages of the Vp-GNc-HyNA system is using a tumor cell-targeted nanocarrier based on Hy as a nanoscaffold. Since Hy can actively target the Hy-receptor, CD44, which is overexpressed on various types of tumor cells, HyNA could also readily target tumors by an active targeting mechanism.19,36−38 To test tumor targetability of the nanoassemblies in vitro, we examined cellular uptake of GNc-HyNA under confocal microscopy. GNc-HyNA labeled with the NIR fluorescence dye Cy5.5 was incubated with CD44-positive MDA-MB-231 breast cancer cells or CD44-less-positive NIH3T3 cells. Strong fluorescence signals from GNc-HyNA were observed in the cytosol of MDA-MB-231 cells, but only faint signals were detected in the NIH-3T3 cells (Supporting Information Figure S6). When we pretreated MDA-MB-231 cells with an excess amount of free hyaluronan molecules to block CD44 receptors, cellular uptake of GNc-HyNA was also significantly inhibited (Figure 3A). From these results, we confirmed that the uptake of GNc-HyNA into tumor cells was mainly attributable to the active interaction between hyaluronan and CD44 receptors on the cells. Although gold nanoclusters were installed on the surface of the nanoassemblies, Hy polymeric chains were still exposed on the surface of GNc-HyNA. The negative surface charge and CD44mediated endocytosis of GNc-HyNA substantiated that (Figure 3A, Supporting Information Table S1 and Figure S6). In addition, we also visualized intracellular Vp or gold nanoclusters delivered by Vp-GNc-HyNA. After Vp-GNcHyNA was incubated with MDA-MB-231 cells, noticeable fluorescence signals from Vp and dark-field signals from gold nanoclusters were observed in the cytosol of the cells (Figure 3B). Strong fluorescence signals from Vp in the cells indicate the release of Vp molecules from GNc-HyNA. Intracellular Vp release can mainly be ascribed to enzymatic degradation of the Hy backbone by the intracellular enzyme hyaluronidase, known to be overexpressed in various cancer cell lines.39,40 It is also attributed to the dissociation of the gold shell from Vp-GNcHyNA given the strong Au−S interaction (∼40 kcal/mol) between gold nanoclusters and intracellular glutathione (GSH) molecules, which can replace the Au−N interaction (∼8 kcal/ mol).41,42 To scrutinize the intracellular drug release mechanism, we monitored drug release profiles in the presence of two intracellular enzymes, GSH and hyaluronidase (Hyal). Incubating Vp-GNc-HyNA with GSH and Hyal significantly accelerated Vp release (Figure 3C). In PBS,
