Cetyltrimethyl ammonium bromide (CTAB), tetraethoxysilane (TEOS), and Hoechst 33342 were purchased from Sigma-Aldrich (St Louis, United States). Indocyanine Green (ICG) and 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine (DiD) were purchased from Absin (Shanghai, China). Human cervical cancer cells (Hela) were purchased from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences (Shanghai, China). The membrane protein extraction kit, calcein-AM/PI Double Staining Kit, Cell Counting Kit-8 (CCK-8) and penicillin–streptomycin were purchased from Biyuntian (Jiangsu, China). Dulbecco’s modified Eagle medium (DMEM) was purchased from Gibco (Grand Island, United States). Fetal bovine serum (FBS) was purchased from Hyclone (Logan, United States). All other reagents were of analytical grade without any purification.

MSNs were prepared using a sol–gel method according to the previous studies with minor modification (Goel et al., 2014; Quan et al., 2015; Chaudhary et al., 2019). In brief, CTAB (0.5 g) and NaOH (0.14 g) were dissolved in distilled water (250 ml) and stirred at 80°C for 30 min. Then, the silica precursor TEOS (2.5 ml) was added dropwise to the CTAB/NaOH mixture and stirred at 80°C for 2 h. The MSNs were collected as pellets by centrifugation (20,000 g, 15 min) and washed with ethanol three times. To remove the surfactant CTAC, MSNs were extracted with a 1 wt% solution of NaCl in ethanol at 40°C for 24 h thrice. Finally, MSNs were resuspended in PBS and stored at 4°C for further use.

ICG was loaded onto the MSNs using a simple incubation method (Chaudhary et al., 2019). To maximize ICG loading of MSNs, 0.2 mg of ICG was added to the MSN suspension (1 mg/ml, 1 ml), and the mixture was stirred magnetically at room temperature for 24 h. Then, the ICG-loaded MSN nanoparticles (MSNs@ICG, MI) were collected by centrifugation (20,000 g, 15 min) and washed with PBS three times. Finally, MI was redispersed in PBS (pH 7.4).

MD was prepared for the optimization of cellular uptake. DiD was dissolved in DMSO and consequently diluted with PBS. Then, 0.2 mg of DiD was loaded onto MSNs (1 mg/ml, 1 ml) by the same incubation method mentioned before.

Cancer cell membrane–coated MSNs@ICG (CMI) was prepared according to the previous report (Fang et al., 2020). In brief, the human cervical cancer cells, Hela, were maintained in DMEM supplemented with 10% FBS, 100  U/mL penicillin G, and 100 μg/ml streptomycin. And the cells were cultured at 37°C in a humidified atmosphere of 5% CO₂. The cancer cell membrane (CCM) was extracted from human cervical cancer cells Hela by using a membrane protein extraction kit, following the instructions from the manufacturer (Biyuntian, China). The CCM was added to the MI dispersion, and the mixture was ultrasonically dispersed by a Scientz-IID ultrasonic homogenizer (Ningbo Scientz Biotechnology Co., Ltd., China) for 1 h. Afterward, CMI was extruded by a mini-extruder (Avanti, Canada) through the 100-nm polycarbonate membrane 20 times in the dark. The CCM coating of MSN@DiD (CCM/MSNs@DiD, CMD) was also prepared by the same procedure mentioned before.

Cervical cancer Hela cells were maintained in DMEM supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/ml streptomycin. Hela cells were cultured at 37°C in an incubation of the humidified atmosphere with 5% CO₂.

In order to confirm the homologous targeting ability of cancer cell membrane–coated MSNs, DiD was chosen as a model fluorescent dye, and cellular uptake of CMD with different concentrations of DiD was visualized by confocal microscopy (Zeiss, Germany). In brief, Hela cells (1×10⁵ cells/well) were seeded in 6-well plates and then cultured for 24 h. Hela cells were exposed to CMD with various DiD concentrations (1.25, 2.5, 5, and 10 μg/ml), respectively. After a 12-h incubation, the treated Hela cells were washed three times with cold PBS. And, cell nuclei were stained with Hoechst 33342 for 15 min. The cellular uptake images were visualized by a Zeiss LSM 880 confocal laser scanning microscope (Zeiss, Germany).

Six-week-old female BALB/C nude mice (20 g weight) were purchased from the Shanghai SLAC Laboratory Animal Co. Ltd (Shanghai, China). All the animal experiments were approved by the Institutional Animal Care and Use Committee of Guangzhou Medical University.

In vivo PA imaging was studied on the Hela tumor-bearing mice. In brief, 2×10⁶ Hela cells per mouse were transplanted subcutaneously into the right flanks of female nude mice. When the tumor volume reached about 100 mm³, the mice were randomized into three groups.

In PA imaging, the mice were anesthetized with 5% isoflurane. Then, free ICG, MI, and CMI were intravenously injected via the tail vein with an ICG dose of 10 mg/kg, (n = 3/group) (Jiang et al., 2015; Ma et al., 2021). The real-time tumor PA imaging and PA signal of free ICG and ICG formulations were recorded at 48 h postinjection using the Vevo LAZR-X multimode imaging system.

In order to further confirm the homologous targeting ability, PA imaging was also carried out via the Hela tumor-bearing mice model with the tumor cut in half or completely, followed by treatment with free ICG, MI, and CMI. In brief, for the mice with the tumor cut in half, the Hela tumor-bearing mice were anesthetized with 5% isoflurane, and then the tumor in mice was partially excised. And for the mice with the tumor cut completely, the mice were anesthetized with 5% isoflurane, then the tumors in mice were completely resected. The treated mice were injected intravenously with free ICG, MI, and CMI via the tail vein at an ICG dose of 10 mg/kg (n = 3/group). The real-time tumor PA imaging and PA signal of free ICG and ICG formulations were recorded at 48 h postinjection.

All the results were expressed as mean ± SD. And the statistical analysis was performed in SPSS 13.0 software (SPSS, United States) via one-way ANOVA with Tukey’s post hoc test. A p value < 0.05 was considered statistically significant.

The preparation of CMI is shown in Figure 1. MSNs were prepared by a sol–gel method, in which the surfactant CTAC was used as a template and TEOS was chosen as a silica precursor. Then, ICG was loaded onto the MSNs by a simple incubation method. To construct CMI, human cervical cancer cells Hela as the cancer cell membrane (CCM) was extracted. And the CCM was coated onto the surface of MI by ultrasound vibration and physical extrusion.

The TEM images showed that MSNs, MI, and CMI display spherical shape with a diameter of approximately 50 nm (Figure 2A). As measured by dynamic light scattering (DLS), the hydrodynamic size of the MSNs was 159 ± 15 nm with a polydispersity index (PDI) of 0.13 (Figure 2B). Similarly, the hydrodynamic size of MI was 171 ± 15 nm, with a PDI of 0.14, which indicated that the ICG loading has no significant effect on the MSN particle size (Figure 2B). After CMM coating, the particle sizes of CMI were approximately 200 nm, with a PDI less than 0.15 (Figure 1A). The particle size of MSNs has a significant impact on the passive targeting efficiency of tumors. Mesoporous silica nanoparticles, whose particle size is smaller than 300 nm, might enhance a significant tumor accumulation via the EPR effect (Wang et al., 2015). The prepared CMI was small with narrow size distribution and good dispersion. These results indicated that CMI might be a promising nanoprobe for passive tumor targeting and PA imaging. The zeta potential of MSNs, MI, and CMI was approximately −30 mV (Figure 1B), which might also display good stability and long circulation characteristics for PA imaging (Cheng et al., 2020; Fang et al., 2020; Li et al., 2020). ICG possessed slight negative charges (Lajunen et al., 2018). Therefore, the driven force of MI might be mainly affected by the negative-charged MSNs. Previous reports indicated that the steric hindrance of the glycosylated domain produced the driving forces and guided the correct protein right-side-out orientation by the reduced energetic profile (Hu et al., 2014; Lin et al., 2021). Herein, CMI showed similar zeta potential to MI after CCM coating. To evaluate the stability, the particle sizes of MI and CMI were determined by DLS. CMI showed higher stability within 72 h than MI, which might be beneficial with the CCM coating.

To confirm the cell membrane modification on MSNs, the SDS-PAGE assay was carried out. The SDS-PAGE analysis showed that CMI presented a protein profile very similar to that of Hela cell membrane lysates, indicating that the cell membrane was successfully coated onto the surface of MSNs (Figure 2C). The standard curve of ICG was measured using the UV-vis method (Figure 1D). The ICG encapsulation efficiency of MSNs was about 76%, and the ICG loading efficiency was about 13% when ICG concentration was 0.2 mg/ml (Supplementary Figure S2).

It was expected that the CCM coating from Hela cells would endow the homologous targeting ability to cervical cancer cells. To investigate the homologous targeting ability of CMI, cellular uptakes of CMI were observed in Hela cells via confocal microscopy. Considering that ICG is an NIR dye which effectively absorbs NIR wavelength of about 800 nm and might not be suitable for confocal microscopy, DiD was chosen as a fluorescent probe instead. Cellular uptake of the CMD with different DiD concentrations was visualized by confocal microscopy.

As shown in Figure 3, the DiD fluorescent intensity was found in Hela cells but not overlaid on the Hoechst nucleic acid staining, which indicated that the CMD might be distributed in the cytoplasm of Hela cells after internalization. Furthermore, with the CMD concentration increased, the DiD fluorescence intensity of Hela cells was found to be increased. The phenomenon demonstrated that the cellular uptake of the CMD in Hela cells exhibited a CMD concentration–dependent behavior. The cellular uptake of the CMD showed the most significant fluorescent intensity when DiD concentration reached upto 10 μg/ml. However, the bare nanoparticles without the CM showed lesser DiD signals (Figure 3). The confocal microscopy images indicated that the CMD might contribute the homologous targeting ability to cervical cancer Hela cells with CCM coating (Xu et al., 2020). To further confirm the homologous targeting ability, in vivo PA imaging should be carried out via Hela tumor-bearing mice with the tumor cut in half or completely.

The fluorescent nanoprobe for application should be biocompatibility and low toxicity. In order to confirm the biocompatibility and cytotoxicity of CMI, calcein-AM/PI staining and CCK-8 were carried out. As mentioned before, the results in the confocal imaging showed that the cellular uptake of the CMD in Hela cells exhibited a CMD concentration–dependent behavior. And at 10 μg/ml DiD, the cellular uptake was obviously observed. Therefore, calcein-AM/PI staining was carried out with an ICG concentration of 10 μg/ml, in which the MSN concentration was equivalent with that of the CMD. And the ICG concentration ranging from 0 to 10 μg/ml was performed in CCK-8.

To confirm the biocompatibility of CMI, Hela cells exposed to CMI were stained with calcein-AM/PI for live (green) and dead (red) cell imaging, respectively (He et al., 2020). And the fluorescent images are shown in Figure 4A. Both strong green fluorescent signals and weak red fluorescent signals in Hela cells could be observed after different treatments. And the green and red fluorescent signals of MSNs, MI, and CMI were similar to those of the negative control. These results showed that Hela cell viabilities were not significantly altered after 48-h incubation with MSNs, MI and CMI. However, the red fluorescent signals of free ICG were stronger than those of the negative control and the MSN-based nanoparticles. The fluorescent images indicated that both free ICG and MSN-based nanoparticles showed no apparent toxicity in 48 h. The biocompatibility of MSN-based nanoparticles might be better than that of free ICG.

The in vitro cytotoxicity of ICG and MSN-based nanoparticles was also evaluated using the CCK-8 assay. The cell viabilities of free ICG and MSN-based nanoparticles on Hela cells for 48 h are represented in Figure 4B. The cell viability of MSNs without ICG was above 90%, which indicated that MSNs exhibited good biocompatibility for nanoprobe delivery. The cell viabilities of free ICG declined from 96 ± 4% to 41 ± 3% as the ICG concentration increased to 10 μg/ml. Meanwhile, at the same ICG concentration of 10 μg/ml, the cell viabilities of MI and CMI were 70 ± 6% and 90 ± 2%, respectively. The CCK-8 results were shown to be consistent with those obtained in the calcein-AM/PI dual staining assay.

The cellular uptake and intracellular drug release behaviors in Hela cells from free ICG and CMI were different (Li et al., 2017). In brief, free ICG was a small-molecule probe, with a molecular weight of 774, which entered Hela cells with passive diffusion and was directly distributed in the cellular cytoplasm. CMIs were nanoparticles with a diameter of approximately 200 nm, which entered the Hela cells with cellular uptake and then escaped from endo-/lysosomes into the cytoplasm. Finally, ICG was distributed in the cellular cytoplasm after release from the pores of CMI. It might take more time for ICG to be located in the cytoplasm from the MSNs than that without encapsulation. Therefore, the results from live/dead staining and CCK-8 assay revealed that the cytotoxicity of free ICG was the highest and that of CMI was the lowest in all the ICG formulations, which might imply that the MSN loading and CCM modification might improve the biosafety of ICG.

In this study, CMI was supposed to exhibit the homologous targeting ability to Hela cancer cells in vivo and could be used as a fluorescent nanoprobe of PA imaging for cervical cancer. To verify the PA imaging effect, in vivo PA imaging of CMI on the Hela tumor-bearing mice was evaluated, and the PA signals were obtained after intravenous injection.

It is reported that ICG with an LD50 value of 50–80 mg/kg exhibits low toxicity for animals (http://www.drugs.com/pro/indocyanine-green.html). Doses of ICG ranging from 0 to 10 mg/kg were used to test NIR imaging by syngeneic murine flank tumor models. And it was found that the fluorescence was optimal for NIR imaging at 5 mg/kg to 10 mg/kg for 24 h (Jiang et al., 2015). 10 mg/kg ICG was applied for NIR fluorescence imaging of cervical cancer for 48 h (Ma et al., 2021). Therefore, a dose of 10 mg/kg ICG was chosen in our in vivo PA imaging for 48 h. As shown in Figure 5A, in the CMI group, high-fluorescence signals were easily observed in the tumor sites. The PA intensities gradually increased in the tumor site, then reached its maximum at 12 h, and finally persisted up to 48 h after injection. Further semiquantitative analysis (Figure 5B) showed that the PA intensities of both MI and CMI in the tumor were significantly stronger than those of free ICG. The PA intensity of CMI in the tumor site was 76% at 12 h, which was 0.9-fold higher (p < 0.01) than that of free ICG and 1/3-fold higher (p < 0.05) than that of MI. Results showed that CMI had stronger PA intensity in the tumor site than MI and free ICG, implying that CMI could specifically accumulate into the cervical tumor, which might be attributed to the homologous targeting ability.

According to the homologous targeting ability, CMI could show enhanced PA imaging on the cervical tumor, but the enhanced PA imaging might disappear after complete tumor resection. In order to further confirm the homologous targeting ability, in the following PA imaging experiments, the tumors were cut in half or completely in the mice, followed by treatment with free ICG, MI, and CMI. We found that the PA signals in the half tumors treated with CMI were also much higher than free ICG or MI-treated ones (Figures 6A,B). However, after the tumors were excised, the PA signals in the CMI-treated group recovered to the free ICG or MI level (Figures 7A,B). Interestingly, the in vivo PA intensities of CMI and MI were significantly higher than those of ICG at 12 and 24 h (p < 0.05). But at the same time, the PA intensities of CMI and MI had no significant differences (p > 0.05). The phenomena might be attributed to the MSN encapsulation, which could prolong the circulation of ICG. For PA imaging, free ICG was rapidly cleared in vivo; therefore, the ICG distribution was limited, and the ICG signal was low after 12 h. When encapsulated into the MSNs, ICG was protected from the reticuloendothelial system by MSNs; therefore, the body clearance of ICG was slowed down, and the ICG signal from CMI and MI was higher than free ICG after tumor resection.

These results showed that CMI displayed the strongest PA intensity in the tumor site, demonstrating that CMI could specifically accumulate into the tumor. The superior PA imaging efficacy of CMI might be attributed to the EPR effect via MSN loading and homologous binding ability via CCM coating (Peng et al., 2021).

In our present study, Hela cell membrane coating MSNs was constructed for the homotypic cancer diagnosis. All the results showed that CMI might be a favorable biomimetic nanoprobe for PA imaging, which might be benefited from the MSN loading and Hela cell membrane coating. First, owing to the special structure of MSNs, ICG was successfully loaded into the pore of MSN. And the nanostructure of MSNs could improve the photostability and prolong the circulation of ICG. Besides, MSNs with particle sizes less than 300 nm were endowed with passive targeting ability via the EFR effect. Second, owing to Hela cell membrane coating, CMI was endowed with homologous binding ability to cervical carcinoma, which efficiently delivered ICG to cervical cancer via the homotypic recognition to the same cell lines. Herein, the enhanced PA imaging of cervical cancer was achieved via MSN loading and Hela cell membrane coating.

PA imaging is a new non-invasive imaging technology that combines with ultrasonic detection and optical excitation. The detected tissue acoustic waves which are emitted by the pulsed laser excitation can be detected by an ultrasound transducer, and the PA images are reconstructed by the absorbed optical energy distribution (Zeng et al., 2018). The PA images are locally imaged that generally focus on the detected tissue, which is different from the NIR imaging that could provide with body imaging (Zeng et al., 2018; Chaudhary et al., 2019). In our study, the biomimetic probe CMI was applied to PA imaging on cervical cancer, so the PA imaging focused on cervical tumor sites, and the PA images on cervical tumors were provided. The PA imaging of CMI displayed enhanced PA intensity on the cervical tumor sites compared with free ICG and MI. However, no directing results could be sure whether CMI was penetrated into the tumor or remained on the tumor surface, CMI could improve the fluorescence intensity on the cervical tumor site, which might be beneficial for PA imaging via the homologous targeting ability. Our results showed that CMI could be a promising probe for PA imaging of cervical tumor tissues. And, the multimodality precise diagnosis combined with PA/NIR imaging and photothermal therapy would be performed further, and the total body images and tumor inhibitory effects will be discussed in our future research.