In this experiment, modified CN-RB was synthesized by the thermal copolymerization of melamine and rose bengal (Zhang et al., 2010b). Then, a cell disruptor was used to convert make CN-RB into smaller CN-RB nanoparticles (CN-RB NPs). As indicated by the X-ray diffraction (XRD) spectra (Figure 1C), CN-RB NPs have the same <002> diffraction peak as g-C₃N₄ NPs at 27.6°, which can be ascribed to the stacking of aromatic structures (Zhang et al., 2017). In contrast, g-C₃N₄ NPs shows the <001> diffraction peak at 13.8°, while this does not exist in CN-RB NPs. The <001> diffraction peak is connected with in-plane structure stacking (Liu et al., 2011). The disappearance of the peak at 13.8° in CN-RB NPs is probably due to the change in the stacking structure in the process of copolymerization (Ho et al., 2015). As shown in the X-ray spectroscopy (XPS) spectra (Figure 1D), CN-RB NPs contain four elements of C, N, O and iodine I). The high-resolution image of C 1s (Fig. S1a, ESI†) shows three peaks at 288.3 eV, 285.7 eV, and 284.8 eV. The main peak of C 1s at 288.3 eV is sp² hybrid carbon and the weaker peak at 284.8 eV corresponds to graphite-phase carbon (Liu et al., 2015). The high-resolution images of N 1s and O 1s are shown in Fig. S1b and S1c (ESI†). The peaks of N 1s are 398.8 eV and 399.6 eV, and the peaks of O 1s are 535.6 eV, 532.8 eV, and531.5 eV, respectively. In addition, the high-resolution peaks of I 3 days (Fig. S1d, ESI†) are seen at 630.1 eV and 619 eV (Zhang et al., 2014). The presence of I element can be confirmed from the XPS spectrum, but it does not appear in the XRD, probably due to the low content of I element. Supported by the Fourier transform infrared (FTIR) spectra in Fig. S2 (ESI†), the multiple bands of CN-RB NPs are located at 809 cm⁻¹, 1200–1650 cm⁻¹, and 3400 cm⁻¹, which confirm that the copolymerized CN-RB NPs have the skeleton of the g-C₃N₄ NPs (Ho et al., 2015). In addition, the CN-RB NPs show two resolved signals in the solid-state NMR 13C {1H} cross-polarization spectrum (Figure 1E) at 157 ppm and 165 ppm, which are the same as in g-C₃N₄ NPs. The above results further indicate that the copolymerization of CN-RB NPs leads to an identical chemical skeleton to g-C₃N₄ NPs.

The hydrodynamic diameter of CN-RB NPs in aqueous solution (Fig.S3, ESI†) is approximately 130 nm, which is suitable for biological application (Wu et al., 2016). It can be seen from the transmission electron microscopy (TEM) image (Figure 1A) that the particle size of CN-RB NPs is approximately 100 nm, which is slightly smaller than the size of the hydrodynamic diameter. As shown in Figure 1B, the high-resolution transmission electron microscopy (HRTEM) image suggests that the lattice spacing of CN-RB NPs is 0.340 nm, which indicates that CN-RB NPs have a graphene-like structure (Stankovich et al., 2007). Figure 1F shows that the absorption of CN-RB NPs in the NIR-II window is significantly enhanced, which indicates that CN-RB NPs can be used as a NIR-II window photothermal agent. In order to verify that the enhanced absorption is not solely induced by the carbon-rich molecular structure, the absorption of CN-B NPs is compared with CN-LB NPs synthesized by melamine and rhodamine B, as they have a similar structure to rose bengal. The results (Fig. S4, ESI†) show that the absorption of CN-RB NPs in the NIR region is much stronger than that of CN-LB NPs, confirming that the doping of I element indeed contributes to the increased absorption of CN-RB NPs in the NIR region. The above results all suggest that the obtained CN-RB NPs not only maintain the skeleton of g-C₃N₄, but also enhance the optical absorption in the NIR-II window, which will lay a solid foundation for their application in biomedicine.

The absorbance of CN-RB NPs is extended from visible light to the NIR window (>1000 nm), suggesting that CN-RB NPs have potential for NIR-II PTT. As the absorption of CN-RB NPs has no sharp peaks, we choose the 1064 nm laser to trigger NIR-II PTT. We first tested the photothermal properties of CN-RB NPs at different concentrations (50 ppm, 100 ppm and 200 ppm). Figure 2A shows that under 1064 nm laser irradiation for 15 min, the increased temperature of the CN-RB NPs solution is dependent on the increasing concentration. At the concentrations of 50 ppm and 100 ppm, the temperature of the CN-RB NP solution can increase to 34.3°C and 40.3°C, respectively. When the concentration is 200 ppm, the temperature of CN-B NPs can rise to 46.3°C, which is suitable for mild-temperature photothermal therapy (Liu et al., 2017). Next, the photothermal stability of CN-RB NPs was measured in a cycle experiment. CN-RB NPs solution (200 ppm) was first irradiated with a 1064 nm laser (2 W cm⁻²) for 15 min, and then the solution was allowed to cool to the original temperature (24.7°C). As shown in Figure 2B, after five cycles, the maximum temperature of the CN-RB NP solution did not change in each cycle, indicating that CN-RB NPs have good thermal stability. The photothermal conversion efficiency η) of CN-RB NPs was calculated to be 35%, which is comparable to some nanomaterials with NIR-II mild photothermal effects (Yu et al., 2020). The above results suggest that CN-RB NPs can be used as an NIR-II photothermal agent.

Here, the cell cytotoxicity of CN-RB NPs was appraised with an MTT assay. The cell viability of 4T1 cells incubated with different concentrations (0, 1, 2, 5, 10, 20, 50, 100, and 200 μg mL⁻¹) of CN-RB NPs for 24 h was above 95% in Figure 2C. It indicates that there is no obvious cellular death produced by CN-RB NPs in the dark.

To estimate the in vitro therapeutic efficiency of CN-RB NPs, the MTT assay was performed to determine the cell viability under different experimental conditions. It can be seen from Figure 2D that the survival rate of 4T1 cells treated with X-ray alone and a 1064 nm laser alone was 96% and 98%, respectively. In addition, we considered the radiation therapy of g-C₃N₄ NPs. As shown in Fig. S5, the death rate of cells co-incubated with g-C₃N₄ NPs is approximately 10% after X-ray irradiation, indicating that g-C₃N₄ NPs have a weak response to X-ray. However, when treated with the PTT or RT of CN-RB NPs, the survival rate of 4T1 cells can be reduced to 65% and 62%, respectively, which suggests that CN-RB NPs not only have NIR-II PTT capability, but can be used for RT. With PTT and RT co-therapy, the cell viability decreases to 45%, confirming the in-vitro-enhanced PTT/RT co-therapy. The in vitro cells experiment encourages us to carry out further animal experiments.

Based on the in vitro NIR-II PTT/RT co-therapy of CN-RB NPs, we studied the tumor-suppressive effect of the CN-RB NPs in 4T1-tumor-bearing BALB/c nude mice. In order to evaluate the photothermal effect of CN-RB NPs in the NIR-II window, we used an infrared thermal imager to record the temperature of the tumor site in mice injected with CN-RB NPs under 1064 nm laser irradiation (2 W cm⁻²). It can be seen from Figure 3A that after being irradiated for 15 min, the temperature of the tumor site of the mouse could reach 45 °C, while the temperature of the tumor site of the mice injected with PBS hardly changed. The above data confirm that CN-RB NPs have a good photothermal effect under 1064 nm laser irradiation. PA imaging is widely used in biomedicine due to its high resolution and deep penetration. We first measured the PA signal of the CN-RB NPs solution and pure water from 1200 to 2000 nm. Figure 3B shows that in the range of 1200–1280 nm, the quantified PA signal of CN-RB NPs is 7 times stronger than that of pure water. At the wavelength of 1300 nm, the quantified PA signal of CN-RB NPs is 4 times stronger than that of pure water. CN-RB NPs have a strong PA signal in the NIR-II window. Furthermore, we evaluated the PA signal of CN-RB NPs in vivo in tumor tissues. Figures 3C,D show that, in contrast to the case before injection, the PA signal at 1200 nm and 1280 nm in the tumor tissues injected with CN-RB NPs was increased by 2.7 times and 2 times, respectively, indicating that CN-RB NPs are suitable for NIR-II PA imaging.

To evaluate the therapeutic efficiency in vivo, the mice were divided into five groups (PBS, PBS + RT, CN-RB NPs+1064 nm, CN-RB NPs + RT, and CN-RB NPs+ 1064 nm + RT, respectively, n = 3). The first group (PBS+1064 nm) and the second group (PBS+1064 nm + RT) were the control groups. Mice in the control groups were intratumorally injected with PBS, and then irradiated with 1064 nm and/or X-ray. Mice of the third group (CN-RB NPs+1064 nm) and the fourth group (CN-RB NPs + RT) were intratumorally injected with CN-RB NPs and then irradiated with a 1064 nm laser or X-ray. Mice in the fifth group (CN-RB NPs+1064 nm + RT) were intratumorally injected with CN-RB NPs, first irradiated with a 1064 nm laser, and then irradiated with X-ray. As shown in Figure 4A, there were no obvious loss of body weight or death in all groups, which indicate that CN-RB NPs have no significant toxicity. It can be seen from Figure 4B that, after 14 days, the tumors of the mice in the first and second groups grew rapidly. In contrast, the tumors in the third, fourth, and fifth groups were suppressed. Among them, the tumor inhibition effect in the third group of mice and the fourth group of mice was equivalent, which showed that CN-RB NPs not only had a photothermal treatment effect under 1064 nm laser irradiation, but also had a radiotherapy effect under X-ray irradiation. At the same time, the tumor suppression effect in the fifth group of mice was the most obvious compared with the other four groups, which indicated the enhancement caused by the NIR-II PTT/RT co-therapy of CN-RB NPs. After 14 days, the tumor tissues were removed from the mice. As can be seen from Figures 4C,D, the tumor tissues of mice in the treatment group were much smaller than those in the control groups. The temperature of tumor site of the mouse after 15 min’ photothermal/radiotherapy can reach to 45°C, this may cause a little bit skin burnt after treatment. However, the main purpose is removing malignant tumor (in our experiment, breast cancer) completely. Aesthetic problem is not our primary concern. We think a little superficial burning scar is acceptable. Moreover, among the treatment groups, the therapeutic efficiency of the NIR Ⅱ PTT/RT co-therapy group was the best. We also estimated the therapeutic therapy of CN-RB NPs by TUNEL assay. As shown in Figure 4E, the group that underwent NIR-II PTT/RT co-therapy showed the most apoptotic cells, which indicates that synergistic treatment can achieve a better therapeutic effect. To further consider the side effects of the long-term accumulation of CN-RB NPs, we next estimated the potential toxicity of CN-RB NPs in vivo. After 14 days, the hearts, livers, spleens, lungs, and kidneys of all five groups of mice were collected and stained with hematoxylin and eosin (H&E). There was no obvious pathological damage in all organs (Figure 5), which suggests that CN-RB NPs exhibit low toxicity in vivo and can be used as a long-term therapeutic agent. The all above results demonstrate that CN-RB NPs can be applied in NIR-II PTT/PT co-therapy and have long-term biostability.

Melamine and rose bengal were purchased from Aladdin. DCFH-DA and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide (MTT) were purchased from Sigma-Aldrich. RPMI 1640 medium and Sciecell fetal bovine serum were purchased from Hyclone. Trypsin—EDTA (0.25%), dimethyl sulfoxide (DMSO), penicillin and streptomycin were purchased from Gibco. All chemical reagents were of analytical grade and were used as received without further purification.

CN-RB NPs were synthesized through one-step thermal copolymerization (Zhang et al., 2010b). Briefly, 1.0 g of melamine and rose bengal in a ratio of 1:1 was ground in a mortar for 30 min and then calcined in a muffle furnace at 520°C for 4 h, with a heating rate of 5°C per min. After cooling to room temperature, the obtained sample was in the form of CN-RB. Next, 500 mg of CN-RB was added to 50 mL of ultra-pure water and crushed in an ultrasonic cell breaker in an ice bath for 30 min (power was 500 W, working for 3 s, 6 s apart). The suspension was centrifuged at 8000 rpm for 10 min to remove unexfoliated CN-RB. Finally, the collected supernatant was labelled as CN-RB NPs and was stored in the refrigerator at 4°C for future use.

The morphology and size of the CN-RB NPs were measured through transmission electron microscopy (Talos F200X G2). XRD was performed using an XRD-7000 (Shimazdu, Japan). X-ray photoelectron spectroscopy (XPS) spectra were obtained using an AXIS UltraDLD (Shimazdu, Japan). The UV—vis absorption was measured using a UV-1800 spectrophotometer (Shimadzu, Japan). Photoacoustic imaging was performed using a photoacoustic imager (Fujifilm VisualSonics/VEVO LAZR-X) and X-ray irradiator (Rad Source).

The photothermal conversion efficiency ( η ) was measured by the following Equation 1: (Li et al., 2019) η = h A T 2 − T 1 − Q 1 I 1 − 10 − A λ (1)

h : heat-transfer coefficient;

A : surface area of the container;

T 2 : the equilibrium temperature;

T 1 : the ambient temperature;

Q 1 : the heat dissipation from the quartz cell;

I : laser power;

A λ : the absorbance of CN-B NPs at wavelength of 808 nm.

The h A value was calculated by the following Equation 2 (Ku et al., 2012): h A = m C P τ s (2)

m : the solution mass;

C P : the heat capacity of water;

τ s : the sample time constant.

To evaluate the cytotoxicity of the CN-RB NPs in 4T1 cells, 4T1 mouse breast cancer cells were seeded in a 96-well plate (5×10³ cells mL⁻¹) and were incubated with different concentrations of CN-B NPs (0, 1, 2, 5, 10, 20, 50, 100, and 200 μg mL⁻¹) for 24 h at 37°C and 5% CO₂. After 24 h, 10 µL of MTT solution (5 mg mL⁻¹ diluted in PBS) was added into each well. After 4 h incubation at 37 °C, the culture medium was discarded and 100 µl of DMSO was added to dissolve the formed formazan crystals. The amount of living cells was measured using an enzyme mark instrument at 490 nm.

First, 4T1 mouse breast cancer cells (5×10³ cells mL⁻¹) were seeded in 96-well plates. The medium was replaced after incubation at 37°C in a humidified environment with 5% CO₂, 21% O₂ for 24 h. The experimental sample was divided into eight groups (n = 6). The first four groups were control groups. The first group was the cell control group; the cells in the second group and the third group were irradiated by a 1064 nm laser and X-ray, respectively. The cells in the fourth group were first irradiated by a 1064 nm laser and then X-ray. The last four groups were the experimental groups. The fifth group was only incubated with CN-RB NPs (200 μg mL⁻¹) for 4 h and the sixth group was first incubated with CN-RB NPs (200 μg mL⁻¹) for 4 h and then irradiated with a 1064 nm laser for 15 min. The seventh group was first incubated with CN-RB NPs (200 μg mL⁻¹) for 4 h and then irradiated with X-ray. After incubation with CN-RB NPs (200 μg mL⁻¹) for 4h, the eighth group was irradiated by a 1064 nm laser for 15min and then irradiated by X-ray. After continuing to cultivate for 20 h, 10 µl of MTT (5 mg mL⁻¹) was added to each well for all groups. The medium was removed after 4 h incubation at 37°C and 100 µl of DMSO was added to dissolve the formazan crystals. The absorption at 490 nm was measured using an enzyme mark instrument.

The 4T1 mouse breast cancer cells were inoculated in the right underarm of the 4T1-bearing-tumor BALB/c mice. When the tumor volume reached 100 mm³, 100 µl CN-RB NPs (300  μg mL⁻¹) solution was injected into the tumor tissues of the mice. Then, infrared thermal imaging was performed using an infrared thermal imager at different time points with 1064 nm laser irradiation.

The 4T1 mouse breast cancer cells were inoculated in the right underarm of the mice. When the tumor volume reached 100 mm³, 100 µL of a CN-RB NP solution (300 μg mL⁻¹) was injected into the tumor tissues of the mice. PA imaging of the tumor site was performed using a photoacoustic imager.

Based on the results of in vitro experiments, the therapeutic efficiency of CN-RB NPs was evaluated in 4T1-tumor-bearing BALB/c nude mice. All animal operations were performed under protocols approved by the Animal Experiment Center of Shanghai Jiao Tong University. When the average tumor volume of the mice reached approximately 100 mm³, the mice were randomly divided into four groups (n = 3). The first group and the second group were intratumorally injected with 100 µl of PBS, and the third, fourth and five groups were intratumorally injected with 100 µl of CN-RB NPs (300 μg mL⁻¹). The second group and the fourth group were irradiated only with X-ray. The third group was irradiated only with a 1064 nm laser (2 W cm⁻²) for 15 min. The fifth group was irradiated with a 1064 nm laser for 15 min and then irradiated with X-ray. The concentration of CN-RB NPs was 2 mg kg⁻¹. The body weight and tumor diameter (using a Vernier caliper) were measured every day, and the tumor volume was calculated as V = d²×D/2 (d was the shortest diameter of the tumor, and D was the longest diameter of the tumor).

Graphpad Prism 7.0 statistical software (La Jolla, CA) was used for data analysis, and one-way analysis of variance (one-way ANOVA) was used for comparison among multiple groups. Data are expressed as mean ± standard deviation, p < .05 as well as p < .01 means the difference is statistically significant.