Results

The first stage of the cooperative nanoparticle system, the photothermally-heated gold nanorods, has already been demonstrated by our group (6). Polyethylene glycol (PEG)-coated NRs with a maximum optical absorption of 800 nm are found to accumulate passively in a MDA-MB-435 xenograft tumor. Effective in vivo photothermal heating of the tumor is achieved by application of NIR irradiation (810 nm, ∼0.75 W/cm²) from a diode laser (Fig. 1B).

A cyclic nine-amino acid peptide (Cys-Gly-Asn-Lys-Arg-Thr-Arg-Gly-Cys), referred to as LyP-1, was chosen as the targeting ligand based on a screen of several tumor targeting peptides in MDA-MB-435 xenograft tumors, which showed enhanced LyP-1 accumulation in the heated tumors. The LyP-1 peptide has been reported to selectively recognize lymphatics and tumor cells in certain tumor types and subsequently inhibit tumor growth (21, 22). Recently, it was found that the p32 or gC1qR receptor, whose expression is elevated on the surface of tumor-associated cells undergoing stress, is the target molecule for the LyP-1peptide (23). Thus, we investigated whether the enhanced targeting of LyP-1 relates to upregulation of p32 receptors in the heated tumor.

We first tested the level of p32 expression in MDA-MB-435 xenografts as a function of time postheat treatment. An externally measured temperature of 45 °C was chosen for the laser heat treatment based on a preliminary screen of temperature dependent nanoparticle accumulation. It has been reported that cancer cells are most vulnerable to hyperthermia, chemotherapeutics, or a combined therapy above temperatures of 43 °C (18, 20). Expression of p32 on the MDA-MD-435 tumors was slightly upregulated 6 h after heat treatment, which then returned to almost normal levels 24 h posttreatment (Fig. 1C). Compared with the MDA-MB-435 tumors, less significant changes in the level of heat-mediated p32 expression were observed on C8161 tumors, known as the tumor type that expresses a considerably less amount of p32 compared to MDA-MB-435 tumor (23), over a 24 h period postheating (Fig. S1). Expression of p32 in cultured cells upon heat treatment exhibited a pattern similar to the in vivo xenograft results; the extent of p32 expression on C8161 cells (and cell surfaces) was less than that observed with MDA-MB-435 cells (Fig. S2).

We next investigated the interaction of nanoparticles decorated with LyP-1 peptides with cancer cells upon heat treatment. An optimized formulation of NWs was prepared as previously described (24, 25), and coated with LyP-1 peptides via PEG linkers (∼40 peptides per nanoworm). Significant quantities of the LyP-1 peptide-conjugated NWs (LyP1NWs) were internalized into heated MDA-MB-435 cells relative to unheated cells. In contrast, the C8161 cells displayed lower heat-mediated internalization than the MDA-MB-435 cells (Fig. 1d). The colocalization of p32 receptors and LyP1NW was clearly observed in MDA-MB-435 cells, suggesting that the binding and internalization of LyP1NWs are mediated by p32 receptors on the surface of MDA-MB-435 cells. The lack of interaction of LyP1NWs with C8161 cells is presumed to be due to insufficient availability of p32 receptors on the cell surface (Fig. S2). As expected, control NWs exhibited no interaction in either cell type, regardless of the heat treatment (Fig. S3).

The possibility of selective homing of LyP1NWs to heated xenograft tumors in vivo was then tested. Similar to the in vitro results, targeting of LyP1NWs to heated MDA-MB-435 tumors was prominent relative to unheated tumors, since the ability of LyP1NWs to home to heated C8161 tumors was not significantly different relative to the unheated tumors (Figs. 2 and S4). Histological analysis revealed large quantities of LyP1NWs occupying vessel structures that were not colocalized with the blood vessel stain, consistent with the previously reported affinity of LyP-1 for lymphatics (21). In both types of tumors, most of the observed LyP1NWs were either colocalized with p32 receptors or distributed in the extravascular region of the heated tumors. Additionally, the distribution of control NWs in tumors did not correlate with the p32 receptor distribution, even though significant quantities of NWs were observed in the heated tumors. Furthermore, histological images of tumors for which LyP1NWs were administered immediately after heat treatment were similar to those for which LyP1NWs were injected right before heat treatment (Fig. S5), suggesting that prominent targeting of LyP1NWs on the individual cells of heated tumors can be attributed mainly to their binding to the p32 receptors, not the simultaneous hyperthermia.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Temperature-induced amplification of in vivo tumor targeting. (A) Fluorescence intensity from Cy7-labeled LyP-1-conjugated magnetic nanoworms (LyP1NW) and Cy7-labeled control nanoworms (NW) in MDA-MB-435 tumor as a function of externally applied heat (30 min). Heated at (45 °C) and unheated (37 °C) samples indicated with (+) and (-), respectively. The tissues were collected from the mice 24 h postinjection; NIR fluorescence images use Cy7 channel. * indicates P < 0.05 (n = 3 ∼ 4). (B) Fluorescence image of major organs from the mice in (A). T+, T-, Li, Sp, K, and Br indicate heated tumor, unheated tumor, liver, spleen, kidney, and brain, respectively. (C) Histological analysis of LyP1NW or NW distribution in MDA-MB-435 tumors with (+) or without (-) application of external heat. Green indicates NWs (labeled with Cy 5.5). Cellular stains same as in Fig. 1D, blood vessels stained with CD31 followed by Alexa Fluor® 594 goat antirat IgG. Arrowhead indicates a lymphatic vessel structure displaying a signal from the labeled LyP1NWs. Scale bar is 100 μm. Error bars indicate standard deviations from ≥3 measurements.

Having verified temperature-induced amplification of nanoparticle targeting to tumor cells in vitro and to xenografted tumors in vivo, we next evaluated in vitro photothermal-assisted cytotoxicity of targeted therapeutic carriers. Liposomes constructed from lipids that are not thermally sensitive were prepared and loaded with the anticancer drug doxorubicin (DOX) (26). The LyP-1 peptide-conjugated DOX liposomes (LyP1LPs) displayed greater levels of cytotoxicity toward MDA-MB-435 cells relative to control DOX liposomes (DOX concentration > 10 ugDOX/mL in both experiments). Enhanced cytotoxicity was observed for heat-treated (45 °C) samples, whereas the measured difference in cytotoxicity at 37 °C was insignificant (Fig. 3A and 3B). The increased cytotoxicity of LyP1LPs toward heat-treated cells is ascribed to a combination of hyperthermal chemotherapy and targeting to (upregulated) receptor proteins. Although it was reported that LyP-1 peptide itself has a therapeutic effect (22), the peptide amount on the particles is much less than was needed for the antitumor activity. By contrast, the heat-induced cytotoxicity of LyP1LPs toward C8161 melanoma cells was significantly less pronounced; this is attributed to lower levels of expression of p32 on the C8161 cellular surface and higher resistance to DOX, relative to MDA-MB-435 cells (Fig. S6).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Heat-mediated cytotoxicity of targeted therapeutic nanoparticles in vitro. (A and B) Temperature-induced cytotoxicity of various therapeutic molecule or nanoparticle formulations toward MDA-MB-435 human carcinoma cells by MTT assay. The cells were treated with free DOX, control DOX-containing liposomes (LP), or LyP-1-conjugated, DOX-containing liposomes (LyP1LP) with the indicated concentrations of DOX. Samples incubated at 37 °C (A) or 45 °C (B).

Finally, the therapeutic efficacy of the complete cooperative nanomaterials system was tested on a xenograft mouse cancer model. Twenty-four h posttreatment, targeting efficacy of LyP1LPs was significantly larger in the photothermally engineered tumors than in the normal tumors and than that of control LPs (Fig. 4A and B). The results clearly show that targeted LPs display greater accumulation in the engineered tumors and deliver more encapsulated DOX payload relative to untargeted LPs. By contrast, in the normal (unheated) tumor environment, both LP formulations show relatively low levels of accumulation (Figs. 4A and S7). Additionally, in order to achieve therapeutic effects in the unheated tumor, multiple administrations of relatively high doses of LPs are required (Fig. S8). However, addition of the targeting ligand LyP-1 to the LP formulation slows tumor growth, in accord with previous work (27).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Successful antitumor therapy using cooperative nanosystem, demonstrated in mice bearing MDA-MB-435 tumors. (A) Quantification of in vivo accumulation of DOX in tumors as a function of NR-mediated laser heating of LyP-1-conjugated liposomes (LyP1LP) or control liposomes that contain no targeting peptide (LP). NR + L and NR - L indicate mice containing gold nanorods that were or were not subjected to laser treatment, respectively. Amount of DOX present quantified by fluorescence microscopy to yield a percentage of injected dose per tissue mass. * indicates P < 0.05 (n = 3 ∼ 4). (B) Histological analysis of DOX distribution in tumors from the mice in (A) who were subjected to NR-mediated thermal therapy showing the distribution of nanoparticles (Alexa Fluor® 488 label on control liposome and 5(6)-carboxyfluorescein (FAM) label on LyP-1, Green) and DOX (Red). Nuclei stained with DAPI (Blue). Scale bar is 100 μm. (C) Change in tumor volume of different treatment groups containing bilateral MDA-MB-435 xenograft tumors. 72 h postinjection of gold nanorods (NR, 10 mgAu/kg), mice were injected with a single dose of saline, control liposomes (LP), and LyP-1-conjugated liposomes (LyP1LP). “+H (Hyperthermia)” denotes one of the two tumors in the animal that was irradiated with the NIR laser. The tumor not irradiated is indicated as “-H”. Tumor volumes monitored every 3 d postirradiation. Error bars indicate standard deviations from ≥3 measurements. * indicates P < 0.05 and ** indicates P < 0.02 for +H + LyP1LP sample and all other treatment sets (n = 4 ∼ 6). (D) Survival rate in different treatment groups after a single dose (3 mgDOX/kg) into mice (n = 6) containing single MDA-MB-435 xenograft tumors. Error bars indicate standard deviations from ≥3 measurements.

As mentioned above, hyperthermia in the temperature range ∼43 °C has been shown to selectively damage malignant cells relative to normal cells (18). Similarly, the increased temperature in the tumor produced by NR-mediated photothermal heating slows tumor growth in vivo, although it does not reduce tumor volume (Fig. S9). However, tumors (or tumor cells) whose local microenvironment has been engineered by NR-mediated heating are more vulnerable to attack by therapeutic nanoparticles (Fig. 4C and 4D). Combined with NR-mediated photothermal engineering, a single injection of therapeutic nanoparticles at a relatively low therapeutic dose (3 mgDOX/kg) is able to achieve significant tumor regression or elimination, which has not been observed in this tumor model with previous targeted therapies even with multiple high doses (27, 28). For all the treatments studied in this work, no significant loss of body mass was observed.