Preparation and characterization of Au@C-CCM

As to prepare the Au@C-CCM NPs, Au@C NPs were synthesized using perchlorate gold (HAuCl4·3H2O) as the gold source and glucose (C6H12O6) as the carbon source via a one-step hydrothermal method. After Au@C NPs preparation, the membrane of patient-derived cells (PDC) was extracted via a repeated freeze-thaw process, and then coated on the surface of Au@C NPs via sonication-absorption to form Au@C-CCM. The CAL27 CMs were employed as the model CCM. The morphology of Au@C and Au@C-CCM was investigated by transmission electron microscopy (TEM) (Fig. 1a, b). The spherical Au@C has a designed core-shell structure with a dynamic diameter of approximately 150 nm. The CCM coating did not change the spherical morphology of NPs, which formed a clear organic layer on the surface of Au@C, resulting in a larger size of approximately 250 nm for Au@C-CCM (Fig. S1). The distinct size distribution of these NPs and zeta potential are also presented in Figs. S1, S2. The CCM coating process was investigated by Fourier transform infrared (FTIR) spectroscopy. The spectroscopy of Au@C after CCM coating presented additional peaks of C–H (CH2) vibration at 2922 cm−1 and P–O vibration at 1061 cm−1, which could be attributed to the phospholipids in CCM,29 indicating the successful synthesis of Au@C-CCM (Fig. S3). More evidence for the formation of the designed Au@C-CCM was provided by the elemental mapping images of TEM, which showed the characteristic elements of Au, C, and P from the gold core, carbon shell, and CCM corona, respectively (Fig. 1c and Fig. S4). The biochemical structure of CCM before and after coating on Au@C as investigated by Western–Blotting analysis using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie blue staining as well as identification of specific membrane markers. The Au@C-CCM27 showed similar protein banding patterns as the CAL27 CMs and also possessed these specific membrane markers (Fig. 1d and Fig. S5).

Fig. 1

Physicochemical characterization of Au@C and Au@C-CCM.TEM images of a Au@C NPs and b Au@C-CCM NPs. Inset in the upper right corner: TEM image of a single Au@C-CCM NP, scale bar = 50 nm. c EDS elemental mapping of Au@C-CCM. d SDS-PAGE and e vis-NIR absorption spectroscopy of Au@C-CCM dispersions of the indicated concentrations. f Photothermal heating curves of the indicated Au@C-CCM dispersions under 808 nm laser irradiation (1 W·cm−2) for 5 min. g Plots of temperature change (ΔT) versus the concentration of Au@C-CCM. h Temperature change curves of Au@C-CCM dispersion under different laser power densities. i Temperature change curve of Au@C-CCM through five on-off laser cycles. j Photothermal temperature change curve of 100 mg·L−1 Au@C-CCM and water

Au/carbon nanocomposite materials are well-known for their excellent photothermal performance, which can be reflected by their efficient absorption in the near infrared (NIR) region.30,31,32,33 Therefore, the absorption of Au@C-CCM was assessed using UV-visible spectroscopy, which showed a broad NIR absorption band as well as a liner relationship between the absorbance at 808 nm and NP concentration (Fig. 1e and Fig. S6a, b). This strong absorbance at 808 nm was expected to induce the desired photothermal conversion effect, which was evaluated by a thermal IR imager as a temperature detector. We also evaluated the photothermal heating curves for Au@C-CCM and water under 808 or 980 nm laser irritation. The maximum temperature for 200 mg·L−1 Au@C-CCM for 5 min under 808 nm was 66.8 °C, higher than the maximum temperature under 980 nm. The temperature of water increased by only 1.8 °C after 5 min irradiation by an 808 nm laser, while a 980 nm laser led to a temperature elevation of 17.8 °C (Fig. S6c, d). The photothermal performance of Au@C-CCM showed obvious dependence on the NP concentration, radiation time, and radiation power, which was able to accommodate a 49.6 °C temperature increase under proper conditions (Fig. 1f–h and Fig. S7). Moreover, Au@C-CCM also exhibited significant stability in photothermal performance after five cycles of laser on-off experiments (Fig. 1i), which demonstrates the possibility of multiple PTT sessions during the practical application of tumor treatment. The significant temperature increase could be attributed to the high photothermal conversion efficiency of Au@C-CCM NPs, which was calculated to be 44.2% (Fig. 1j and Fig. S8).

In vitro cytotoxicity and homologous targeting performance of Au@C-CCM

We next evaluated the cell viability and homologous targeting effect of the Au@C-CCM NPs on the corresponding tumor cells. Herein, three types of Au@C-CCM NPs (i.e., Au@C-CCM27, Au@C-CCM7, and Au@C-CCM6) were prepared through cloaking Au@C with CCMs from three different HNSCC-tumor-derived cell lines (CAL27, SCC7, and HN6). A cell counting kit-8 (CCK-8) assay was used to show the cell viability of CAL27 cells treated with Au@C or Au@C-CCM27 in the presence/absence of laser irradiation (Fig. 2a). In the absence of laser irradiation, both Au@C and Au@C-CCM27 showed negligible deleterious effects at concentrations of 0–100 mg·L−1, suggesting good biocompatibility. Upon exposure to an 808 nm NIR laser for 10 min at a power density of 1 W·cm−2, we observed enhanced cell death with the increased concentration of both Au@C and Au@C-CCM27. When the NP concentration reached 50 mg·L−1, cell viability decreased to 60.58% and 44.41% for the Au@C- and Au@C-CCM27-treated groups, respectively. Au@C-CCM7 and Au@C-CCM6 also manifested similar photothermal toxicity trends with SCC7 cells and HN6 cells, respectively (Fig. S9), indicating the in vitro PTT effect of the Au@C and Au@C-CCM NPs. The pure CCM also did not influence the cell proliferation and cell viability in these HNSCC cells (Fig. S10).

Fig. 2

In vitro cytotoxicity assays and homologous targeting performance of Au@C-CCM. a Viability of CAL27 cells in the presence/absence of laser irradiation after 24 h incubation with Au@C and Au@C-CCM27 at different concentrations. b Calcein-AM and PI staining of CAL27 cells with different treatments. c Representative CLSM images of CAL27 cells treated with Au@C-CCM27, Au@C-CCM7, or Au@C-RBC. d Flow cytometry and e relative fluorescence intensity of CAL27 cells incubated with Au@C-CCM27, Au@C-CCM7, or Au@C-RBC for 4 h. Data are expressed as the mean ± SD (n = 3). Statistical significance: **P < 0.01, ***P < 0.001

Calcein acetoxymethyl ester (calcein-AM) and propidium iodide (PI) staining assays were conducted to visually distinguish live (green) and dead (red) cells, further demonstrating the antiproliferative effects of Au@C and Au@C-CCM based PTT (Fig. 2b). CAL27 cells mainly remained alive through cocultivation with either Au@C or Au@C-CCM27 at concentrations of 50 mg·L−1 without laser (Fig. S11). In contrast, significant cell death was observed in the two groups in the presence of laser irradiation under the same NPs concentration, confirming the photothermal-induced damage to CAL27 cells. A similar photothermal toxicity trend was also observed in HN6 cells and SCC7 cells (Fig. S12).

Through endocytosis experiments, we observed that the green fluorescence signal of Au@C-CCM27-FITC phagocytosed by CAL27 cells increased over time (Fig. S13). Thereafter, we examined the homologous targeting performance of Au@C-CCM in various cell lines to assess personalized customization potential. We prepared three types of biomimetic nanomaterials coated with different CMs from CAL27, SCC7, or mouse red blood cells (RBC) and labeled them with FITC, yielding labeled Au@C-CCM27, Au@C-CCM7, and Au@C-RBC. Then, CAL27 cells were cocultured with these three Au@C-CCM NPs to assess cell uptake as driven by homologous targeting ability. Confocal laser scanning microscopy (CLSM) images indicated significantly increased phagocytosis in the Au@C-CCM27 NPs as compared to Au@C-CCM7 and Au@C-RBC NPs over the observed time period (Fig. 2c). We also used flow cytometry to show that the Au@C-CCM27 group produced the strongest fluorescence intensity compared to the other groups (Fig. 2d, e), thus confirming the homologous targeting ability of Au@C-CCM27 to the corresponding cancer cells.

In vivo tumor inhibition efficacy of Au@C-CCM in CDX models

To evaluate the ability of Au@C-CCM NPs to inhibit tumor growth after PTT in vivo, Au@C-CCM NPs were prepared using the CAL27 cell line and applied to CDX mice bearing CAL27 tumors. Tumor-bearing mice were randomly divided into four groups, including a group with no treatment (control), a group treated with a laser (laser only), a group treated with Au@C-CCM (Au@C-CCM only), and a group treated with both Au@C-CCM and a laser (Au@C-CCM + laser). Mice in the Au@C-CCM only and Au@C-CCM affiliated laser irradiation groups received 6 mg·mL−1 Au@C-CCM (200 µL) through intravenous injections. The experiment used an 808 nm laser with a power density of 1 W·cm−2, and an IR thermal camera was applied for thermal imaging and recording the temperature changes of the mice upon laser irradiation. We discovered that the temperature of the tumor region increased to 65.8 °C after 20 min of laser irradiation in mice treated with Au@C-CCM, higher than the tumor temperature of mice in the laser group (Fig. 3a and Fig. S14). The results indicated that Au@C-CCM could accumulate into tumors and transfer photo-energy into heat within tumor sites in vivo.

Fig. 3

In vivo photothermal anti-cancer performance of Au@C-CCM in CAL27 tumor-bearing CDX models. a Thermal images of mice in groups treated with/without Au@C-CCM and 808 nm NIR laser irradiation at various time intervals. b Tumor volume change curves following varied treatments (n = 3). c Histology slices of H&E, TUNEL, and Ki-67 staining of different groups; the corresponding magnification images are shown in the lower right corner. d Several randomly selected fields of view under 40× microscope were for quantitative analysis. Data are expressed as the mean ± SD. Statistical significance: *P < 0.05, ***P < 0.001

After various treatments, we subsequently recorded body weight and tumor sizes of mice during a 20 day observation. The body weight change curves indicated no statistical loss in body weight among all groups and demonstrate the biosafety of Au@C-CCM and its PTT (Fig. S15). The tumor volume curves (Fig. 3b) revealed that the tumor sizes in the Au@C-CCM-applied laser irradiation group were significantly decreased. In contrast, the untreated mice had a 3.8-fold increase in tumor volume. We further evaluated the therapeutic efficacy via histological staining of tumor tissues, including H&E staining, Ki-67 immunohistochemical (IHC) staining, and triphosphate nick end labeling (TUNEL) immunofluorescent staining (Fig. 3c). For the Au@C-CCM + laser group, the histology of the tumor region indicated significantly increased necrotic lesions, TUNEL-positive cells, and decreased Ki-67 positive cells when compared to the other three groups. The Ki-67 positive percentages were calculated to be 17.0% in the PTT group, 56.04% in the laser group, 60.7% in the Au@C-CCM group, and 67.84% in the untreated group (Fig. 3d). These results suggest that Au@C-CCM based PTT had an inhibitory effect on tumor cell proliferation and was pro-apoptotic.

Antitumor performance of Au@C-CCM in TOX models

HNSCC which occurs at oral mucosal sites often influenced by complicated oral microbiota, saliva, gingival crevicular fluid, complement, salivary amylase, and other components, which cannot be simulated by the subcutaneous environment.34 Previous studies have shown that orthotopic tumor models can extend the translational relevance of results compared with the subcutaneous milieu.35,36 Therefore, we primarily inoculated HNSCC cells (HN6-luciferase cells) submucosally into the left border of the tongue to simulate the tumorigenicity of tongue cancer.37 Then, Au@C-CCM6 derived from the CM of HN6-luciferase was prepared and intravenously injected 24 h after the inoculation of tumor cells. The fluorescence signals were acquired, and the tumor areas were laser irradiated with a low power density of 0.72 W·cm−2 for 5 min after another 24 h. Through the IR thermal images and temperature curves of the mice, we noted that the maximum temperature in mice that adopted Au@C-CCM6 with laser irradiation increased to 47.1 °C (Fig. 4a, b), a temperature that could induce cell damage.38

Fig. 4

Photothermal anti-cancer performance in orthotopic tongue tumor mouse models. a Thermal images of mice after different treatments during various time intervals. b Temperature change curve between two groups during defined irradiation times. c Representative luminescent images of mice in different groups. d Histology slices of H&E and Ki-67 staining of different groups; the representative image of the entire slice is shown in the lower right corner

Tumor signals were present in the tongue of the mice that merely received laser irradiation. While in the mice treated with Au@C-CCM6 plus laser irradiation, fluorescence signals disappeared on the 7th day without relapse for 30 days of observation (Fig. 4c). Moreover, considering the possible difference in the tumor growth pattern between the orthotopic tumor model and the subcutaneous tumor model,39 we harvested tongue tissue and stained tumor sections to inspect the tissue morphology. H&E staining demonstrated that tumor cells grew unrestrictedly and disorderly under the mucosa of the tongue in the control mice and mice treated with laser irradiation. In mice treated with Au@C-CCM6 based PTT, the tongue surface returned to a soft texture in gross feature and normal epithelial morphology with laminated cells, and no typical tumor cells could be observed. Ki-67 staining showed that Au@C-CCM6 based PTT could inhibit tumor cells proliferation in tongue tissue (Fig. 4d and Fig. S16). These results indicated that Au@C-CCM could homologously target orthotopic tongue tumor cells yeilding sufficiently accumulate, and induce photothermal ablation of tumor cells under proper conditions, even though the tumors were nonpalpable in the TOX models.

Evaluation of Au@C-CCM-Mediated PTT induced ICD in vitro and in immune-competent primary and distant tumor models

Au@C-CCM-mediated PTT-triggered ICD was evaluated by the detection of ICD markers, including High Mobility Group Box 1 (HMGB1), calreticulin (CRT), and adenosine triphosphate (ATP). Au@C-CCM with 808 nm laser irradiation induced higher ATP secretion in CAL27, SCC7, and HN6 cells as compared to the other groups (Fig. 5a). Western blotting showed the lower expression of HMGB1 in Au@C-CCM-mediated PTT group than the other groups (Fig. 5b). Moreover, CLSM images exhibited the HMGB1-associated green fluorescence dissipated from the nucleus and CRT exposure of CAL27 cells in Au@C-CCM-mediated PTT group (Fig. 5c, d). A similar trend was also displayed in SCC7 cells and HN6 cells (Figs. S17–S20).

Fig. 5

Au@C-CCM-mediated PTT induces immunogenic cell death in vitro. a Extracellular ATP levels of CAL27 cells, SCC7 cells, and HN6 cells (n = 3). b Expression of HMGB1 protein in CAL27 cells after treatment with Au@C-CCM-mediated PTT. c Confocal microscopic images of the release of HMGB1 in CAL27 cells after varied treatments. d Confocal microscopic images of exposure of CRT in CAL27 cells after varied treatments. Statistical significance: ***P < 0.001

Next, to further verify synergistic Au@C-CCM based PTT and immunotherapy effect in vivo, we inoculated SCC7 cells to the tongue and subcutaneous sites of C3H mouse to simulate primary and distant HNSCC tumors, respectively. Nine days after SCC7 cell inoculation (when the subcutaneous tumor volumes approximately reached ~100 mm3), mice were divided into four groups including the control group, a group administered with programmed death-ligand 1 antibody (aPDL1), a group treated with Au@C-CCM7 based PTT, and a group treated with PTT combined aPDL1. Schematic illustration of treatment procedures was depicted in Fig. 6a. The results demonstrated that both the PTT and PTT combined aPDL1 groups exhibited superior photothermal performance (Figs. S21, S22). The subcutaneous tumor volumes of mice in the control group did not demonstrate a rapid growth tendency due to the nutrition intake limitation attributed to the fast-growing orthotopic tongue tumor. PTT combined aPDL1 treatment significantly inhibited both the subcutaneous and the orthotopic tumor growth compared to the other three groups (Fig. 6b, c), illustrating that the tumor inhibition efficacy in the combined treatment group was significantly better than the monotherapy groups. The survival rate of the PTT combined aPDL1 group was the highest among all groups during 30 days of observation (Fig. 6d). Histological staining of tongue tumor tissues demonstrated that PTT combined aPDL1 treatment induced larger areas of necrosis, a significant decrease of Ki-67-positive proliferating cells, and higher cell apoptosis rate than the other groups (Figs. S23, S24). IHC staining of the ICD markers of HMGB1 and CRT in tongue tumors showed that PTT combined aPDL1 treatment led to the significantly decreased expression of HMGB1 and increased CRT exposure compared with the other groups (Fig. 6e, f). We also harvested the subcutaneous tumor and investigated the infiltration of the immune cells into tumor regions. CD8+ (a marker of CTL) and CD11c+ (a marker of DC) cells were significantly increased in the PTT combined aPDL1 group, as compared to the other groups (Fig. 6g, h). These results demonstrated that the release of damage-associated molecular patterns (DAMPs) led to the recruitment of CTLs and DCs in subcutaneous tumors, further promoting antitumor immunity.

Fig. 6

Au@C-CCM-mediated PTT induces immunogenic cell death in vivo. a Schematic illustration of the timeline for the in vivo animal study. b Subcutaneous tumor volume curves, c orthotopic tumor volume curves, and d survival rates among the four groups (n = 6). e HMGB1 and CRT expression and f quantified results in different groups. g CD8+ and CD11c+ cells expression and h quantified results in different groups. The corresponding magnification images are inserted in the lower right corner (e, g), and several randomly selected fields of view under 40× microscope were for quantitative analysis (f, h). Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001. Schematics were created with BioRender.com

Tumor suppression evaluation of tailored Au@C-CCM in PDX models

PDX models are established by subcutaneously implanting fresh tumor tissues from clinical patients into immune-compromised mice;40,41 these models can preserve the histologic and genetic features of donor tumors after implantation.42 Therefore, we investigated the customized treatment of homologous Au@C-CCM on PDX models. Two HNSCC patients (defined as 2676 and 2939; clinical information in Table S1) were selected. Optical microscopy images of the diversified cell morphology of their primarily cultured cancer cells were shown in Fig. S25. We prepared two biomimetic nanoplatforms coated with CCMs derived from the patients (named Au@C-2676 and Au@C-2939). We also investigated the cell selectivity in Au@C-2676 to SCC2676, demonstrating that SCC2676 phagocytized more Au@C-2676 NPs, thus exhibiting brighter green fluorescence than cells incubated with Au@C-2939 NPs (Fig. S26). Au@C-2676 and Au@C-2939 were then administered to the PDX model established from the patient defined as 2676 at a dose of 6 mg·mL−1 (200 μL) via the tail vein injection, and various treatments similar to those in the CDX models were performed.

As shown in the IR thermal images, the tumor region in the PDX model that received Au@C-2676 was brighter than that in mice which received Au@C-2939, demonstrating higher PTT efficacy of Au@C-2676 (Fig. 7a). We further compared the overall temperature change between the mice that received different treatments (Au@C-2676 and Au@C-2939). The temperature change of the Au@C-2676 group was significantly higher than that of the Au@C-2939 group (Fig. 7b). The constant temperature plots of the tumor region with Au@C-2676 revealed that irradiation caused the temperature to promptly increase by over 10 °C within 1 min and exceed 50 °C within 75 s. In contrast, the tumor region with Au@C-2939 took more than 18 min to reach 50 °C with an identical laser power density (Fig. S27).

Fig. 7

Photothermal anti-cancer performance in HNSCC PDX models. a Thermal images of mice and b temperature change curve of tumors in PDX models (number: 2676) under laser irradiation after intravenous treatment with Au@C-2676 and Au@C-2939 (n = 3). c Tumor volume curves after PTT between different groups during the same observation time intervals (n = 3). d Histology slices of H&E, Ki-67, and TUNEL staining of different groups; The corresponding magnification images are inserted in the lower right corner. e Several randomly selected fields of view under 40× microscope were for quantitative analysis. Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001

Moreover, to test the treatment efficiency of customized Au@C-CCM in PDX models, we measured the tumor volume and body weight of mice among Au@C-2939 with laser group, Au@C-2676 with laser group and the control group for 18 days. The tumor volume curves revealed that tumor growth in the Au@C-2676 group with PTT were significantly inhibited compared to other groups. The control mice had a 4.4-fold increase in tumor volumes. A 2.6-fold increase was observed in the Au@C-2939 group (Fig. 7c). There was no statistical loss in body weight of Au@C-2676 based PTT group. While the body weight of mice in the control group decreased distinctly, probably due to the high tumor malignancy (Fig. S28). These results illustrated that Au@C-2676 resulted in a much more efficient treatment than Au@C-2939, which most likely results from the sufficient accumulation of the homologous NPs within tumors.

For the group treated with Au@C-2676 plus the laser, H&E staining, TUNEL staining, and Ki-67 staining results also demonstrated significant necrosis, cell apoptosis, and cell proliferation-inhibition in Au@C-2676 with laser group compared to the other groups. We also found that Au@C-2676 with laser led to more thorough tumor cells degeneration and more significant interstitial fibrous tissue hyperplasia accompanied by hyaline degeneration than the Au@C-2939 with laser and the control group (Fig. 7d, e).

In vivo biosafety of Au@C-CCM

The biosafety issues of nanomedicines are paramount for their bio-application and clinical translation.43 Au@C-CCM27 was also intravenously injected into healthy mice, and subsequently blood samples were collected and analyzed at different time points. Main organs samples were also collected for histology analysis. Au@C-CCM27 treatment did not produce abnormal indicators of blood biochemistry, including alanine aminotransferase (ALT), aspartate transaminase (AST), blood urea nitrogen (BUN), creatinine (CRE), or uric acid (UA) (Fig. S29). Additionally, H&E staining showed no observable damage to main organs, including the heart, liver, spleen, lung, and kidney (Fig. S30). In order to further evaluate the biosafety of this biomimetic nanoplatform, we prepared CCM vesicles using HNSCC cells and labeled them with FITC to track their metabolic fate in vivo. We acquired fluorescent signals at various time intervals and discovered that the signals decreased and finally disappeared on the 11th day posttreatment (Fig. S31), indicating the elimination of CCM vesicles. These results demonstrated the safety of the biomimetic nanoplatform in vivo to some extent.

Given PDC-CM-camouflaged nanotechnology and customized PTT (PCMPT) is tailored for HNSCC patients, PCMPT could be applied as a promising therapeutic strategy for early stage HNSCC via biopsy sampling. Through HNSCC cells primary culture from biopsy samples, the CMs extraction and cloaking Au@C, the personalized biomimetic NPs based PTT could be administered to patients, and performed as a pre-operative therapy that will effectively control tumor progression or even ablate tumor tissues completely, allowing for reducing the risk of HNSCC development.