2.1 Mechanistic design of UCNPs with switchable NIR emission
In general, the energy transfer rate in a given sensitizer-activator system depends closely on the energy matching between the emission of sensitizer and the absorption of the activator, which shows a decline following an exponential function of the energy gap (Fig. 1a). The upconversion system with an energy mismatch should result in a relatively slow upconversion process than that of the resonantly coupled system (Fig. 1b), allowing to filtrate such upconversion luminescence by reducing the pulse duration time of the excitation laser. Therefore, the dynamic control of switchable upconversion performance in a much simpler structure nanoparticles becomes highly desirable.
To demonstrate the concept-of-proof, we select Yb3+-Tm3+ couple for the NIR upconversion and the Yb3+-Er3+ couple toward the red upconversion. Note that there is an energy mismatch between Yb3+ (2F5/2) to Tm3+ (3H5) while it is resonant for the Yb3+ (2F5/2) and Er3+ (4I11/2). The NaYF4:20%Yb,1.5%Er,0.3%Tm nanoparticles were prepared for the switchable red/NIR luminescence under CW and pulse-wave (PW) excitations (Fig. 1c). These UCNPs show uniform nanoparticles with an average diameter of ∼26 nm according to the transmission electron microscopy (TEM) images (Figure S1a). The adjacent lattice spacing (~ 0.323 nm) calculated by fast Fourier transform (FFT) analysis corresponds to the (111) crystal plane of hexagonal NaYF4 (Figure S1b), which are in well agreement with the selected-area electron diffraction (SAED) patterns (Figure S1c) and X-ray diffraction (XRD) patterns (Figure S2). Figure 1d shows the upconversion emission spectra under 980 nm excitation. As expected, the red emission of Er3+ at 650 nm (UCL650 nm) from its 4F9/2 → 4I15/2 transition and NIR emission of Tm3+ at 800 nm (UCL800 nm) from its 3H4 → 3H6 transition were clearly observed under the CW 980 nm excitation. Intriguingly, upon the pulse 980 nm laser, the UCNPs show a gradually decreased NIR emission with reducing the pulse width to 10 ns, and the emission light intensity ratio of UCL650 nm to UCL800 nm dramatically increases from 0.89 to 8.25, as illustrated in Fig. 1d,e and Figure S3. This observation quantifies the UCNPs an ideal NIR photoswitchable candidate to be used to construct the orthogonal theranostic agents.
To further examine the non-steady-state upconversion mechanism, we measured the time-dependent profiles at 650 nm and 800 nm, respectively. As displayed in Fig. 2a, the red upconversion emission exhibits a faster rise time than that of the NIR emission before reaching the steady-state. This observation suggests that the upconversion of the NIR emission of Tm3+ indeed needs more time by contrast to the red emission originating from the non-resonant energy transfer from Yb3+ (2F5/2) to Tm3+ (3H5). Consequently, the red to NIR emission intensity ratio shows a monotonous decline with the time during the excitation pulse. Hence, reducing the pulse width of the excitation laser become a facial but effective way to switch off the NIR emission. Another merit of this upconversion system lies in the stable color output during a large range of pump power densities (Fig. 2b) and different excitation frequencies (Figure S4), which lays a solid foundation for the subsequent biological application. It should be noted that a co-doping of Er3+ and Tm3+ into NaYF4 lattice imposes no impact on their rise time feature (Fig. 2c), which confirms the dynamic manipulation of the NIR switchable output. The total energy transfer processes during the dynamic control of the excitation are schematically illustrated in Fig. 2d. In fact, the presence of Tm3+ can promote the red upconversion of Er3+ slightly through the energy circling of Er3+ (4I11/2) → Tm3+ (3H5) → Er3+ (4I11/2) (Figure S5). This is in agreement with the faster rise time of the red upconversion emission of the UCNPs with codoping of Tm3+ (Fig. 2e). Taken together, the results clearly confirm the validity of using the non-steady-state excitation towards the remarkable NIR-switchable effect of UCNPs, which has never previously been reported involving the lanthanide-based nanoparticles. In the next, the typical pulse laser with width of 10 ns and 20 Hz was selected for the following phototheranostics investigation.
2.2 Construction and characterization of orthogonal phototheranostic nanoagent
To demonstrate the dynamic control of the orthogonal photodiagnostics and phototherapeutics for the safe imaging-guided on-demand phototherapy, a nanoagent (UCNPs-DI) was synthesized via assembly with a visible-absorbing semiconductor DPP and an NIR-activatable photosensitizer ICG on the surface of the UCNPs (Fig. 3a and Figure S6). The as-prepared DPP possesses high optical absorption extinction coefficient in the visible region of 550–700 nm, ensuring the effective energy transfer from UCNPs to DPP to generate photoacoustic signal via pulsed photoirradiation of the nanoagent. In view of the typical NIR absorption of ICG around 800 nm and the NIR-switchable emission of UCNPs, the photodynamic effect is non-responsive to the short-pulse irradiation of UCNPs-DI, facilitating the safe long-term and real-time photoacoustic imaging. Upon CW 980 nm irradiation, the switching-on NIR emission could activate the ICG to efficiently produce the cytotoxic reactive oxygen species (ROS) for phototherapy. Therefore, the orthogonal control of the designed UCNPs-DI under programmable photoactivation guarantees the safety of imaging and the precision of therapy.
Figure 3b shows dynamic light scattering (DLS) results and transmission electron microscopy (TEM) images of UCNPs and UCNPs-DI nanoagent. On the basis of DLS results, all samples present a narrow size distribution with peaks centered at 26 nm and 106 nm for UCNPs and UCNPs-DI, respectively, confirming a good stabilization of UCNPs and UCNPs-DI. The UCNPs and UCNPs-DI nanoagent seen in TEM could accord with DLS results. Figure 3c presents the absorption spectra of UCNPs, DPP, ICG, and UCNPs-DI. According to the absorption peaks recorded at 650 and 800 nm and the Fourier-transform infrared (FTIR) spectroscopy (Figure S7) of UCNPs-DI, the DPP and ICG were loaded on the nanoagent successfully. The luminescence spectra of UCNPs-DI together with the starting UCNPs were then recorded under the PW/CW excitation of 980 nm laser. As illustrated in Fig. 3d and S8, compared with the starting UCNPs, the visible and NIR emission of UCNPs-DI under CW irradiation was suppressed dramatically due to the energy transfer from UCNPs to DPP and ICG. This was also characterized by the lifetime decline of the 800 nm emission (from 463 to 293 µs, Figure S9).44,45 These results confirmed the successful construction of the aimed phototheranostic nanoagent UCNPs-DI.
To verify the stability of the nanoagent in aqueous solution and physiological fluids, the absorption of UCNPs-DI in PBS solution and blood serum with different concentrations and times were recorded (Figure S10).46 The absorption of nanoagent does not show an obvious change under these conditions. Adjusting the pH scope of the PBS solution from 8.5 to 5.0 also did not affect its absorption and UCL emission under both 980 nm PW and CW excitations (Figure S11 and S12).47 The results show that UCNPs-DI has an excellent consistency in a physiological environment, revealing its high stability under the physiological conditions.
To study the theranostic nanoagent for orthogonal photoacoustic imaging and photodynamic therapy with programmable excitations of CW/PW 980 nm laser, we then evaluated its performance in the photoinduced generation of acoustic effect and singlet oxygen (1O2) in aqueous solutions. Compared with the starting UCNPs, as depicted in Fig. 3e and S13, we observed a strong and concentration-dependent photoacoustic signal (R2 = 0.99359) when the UCNPs-DI irradiated with 980 nm pulsed laser (10 ns, 0.5 W/cm2).48 The PW irradiation of UCNPs-DI had no obvious photothermal effect, while the temperature of UCNPs-DI solution increased dramatically after exposing to the CW irradiation (0.5 W/cm2), which suggests that the UCNPs-DI does not induce hyperthermia damage during the photoacoustic diagnosis (Figure S14a). The high photostability was confirmed by the measurement of the 10-min photoirradiation of the sample with 4 cycles (Fig. 3f and S14b).49 The 1O2 photogeneration of UCNPs-DI was then evaluated using the dye singlet oxygen sensor green (SOSG) as an indicator.50 Fig. 3g and S15a demonstrated that almost no change was observed for the fluorescence of SOSG at 525 nm in the UCNPs-DI solution with PW 980 nm irradiation, revealing no photodynamic toxicity during the photoacoustic diagnosis. On the contrary, 980-nm CW irradiation of UCNPs-DI can generate abundant 1O2, in which the fluorescence of SOSG increased rapidly within 18 min (Figure S15b).
The above results clearly indicate that the UCNPs-DI produced intense photoacoustic signals with negligible photothermal and photodynamic effect upon 980 nm PW irradiation, providing robust ex-vitro evidence for the feasibility of UCNPs-DI as a safe PA imaging candidate for long-time and real-time diagnose or monitoring the therapeutic treatments. On the other side, switching the 980 nm laser into CW modulation activated the significant photodynamic effect to kill cells in lesions through the generation of adequate reactive oxygen species. The overall results provide substantial supports for us to achieve photoacoustic imaging and photodynamic therapy in an orthogonal manner through programmable excitations of UCNPs-DI by CW/PW 980 nm laser.
2.3 Orthogonally regulated target recognition and photodynamic effect in vitro
The cytotoxicity of the UCNPs-DI with PW (10 ns, 20 Hz, 0.5 W/cm2) and CW 980 nm (0.5 W/cm2) irradiation was firstly investigated in detail. The standard MTT assay was performed on human non-small cell human breast adenocarcinoma cancer cells (MCF-7). Figure 4a presents the viability of MCF-7 cancer cells incubated with different concentrations (25, 50, 100, 150, and 200 µg/mL) of PBS (gray), UCNPs-DI (pink), UCNPs-DI + PW (blue), and UCNPs-DI + CW (red). The cell viability was found still over 90% when the concentration of UCNPs-DI up to 200 µg/mL. Thus, the nanoagent should not influence MCF-7 cells viability when their concentration is less than 200 µg/mL. Significantly, PW irradiation of the cells treated with UCNPs-DI did not induce obvious decrease in the cell viability compared with control cells. By contrast, it shows an evident drop in the cell viability under 980 nm CW excitation. When the MCF-7 cancer cells were incubated with 200 µg/mL UCNPs-DI, the cell viability was found dropping to 28 % in the UCNPs-DI + CW group. At this concentration, as shown in Fig. 4b, light dose-dependent cell-killing effect of UCNPs-DI after incubated for 4 h was recorded when exposure to 980 nm CW irradiation. Remarkably, the cell viability of MCF-7 incubated with UCNPs-DI inhibited dramatically to 27.2% with 0.5 W/cm2 980 nm continuous lasers for 3 min.
The intracellular ROS generation in the UCNPs-DI + CW group and ROS non-generation in the UCNPs-DI + PW group were also confirmed using confocal laser scanning microscope (CLSM) as shown in Fig. 4c, which is consistent with the ROS detection results in Fig. 3g. In addition, the cell viability results were also verified by calcein AM (live cells staining; green) and propidium iodide (PI, dead cells staining; red) double-staining, respectively (Fig. 4d). The fluorescence imaging results showed that the UCNPs-DI has negligible cytotoxicity under PW irradiation but high photodynamic cytotoxicity under CW irradiation. We then further evaluated the cell damage mechanism of UCNPs-DI + PW/CW using flow cytometry with Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) staining. The data revealed that in the cells treated with UCNPs-DI and pulsed laser irradiation, 2.67% of cells were stained with Annexin V-FITC and 0.23% cells were stained with PI, suggesting that most cells was undamaged under this condition (Fig. 4e). The measurement taken at the group with UCNPs-DI after 980 nm CW irradiation showed that 33.87% of the cells were stained with FITC and 30.70% of the cells were stained with PI, suggesting that the UCNPs-DI + CW treatment caused obvious membrane damage and massive cell death. In addition, the UCNPs-DI showed negligible inhibitory effects to the normal LO2 cells with or without the 980 nm PW irradiation (Figure S16). The above results confirmed the evidential potentials of UCNPs-DI as safe photoacoustic imaging candidate. Moreover, the results of UCNPs-DI under the 980 nm CW irradiation have validated the nanoagent as a photoswitchable agent for the effective phototherapy.
2.4 In vivo real-time photoacoustic imaging of tumor
With the intense photoacoustic effect and non-photocytotoxicity of UCNPs-DI validated in vitro and in living cells, we then investigated how to use the agent to overcome certain biological drug delivery barriers for the long-term and real-time photoacoustic imaging of specific tumors (Fig. 5a). The PA imaging is a promising diagnostic imaging technique to monitor the molecular distribution and tumor size/morphology for therapeutical guidance due to its high resolution and noninvasive visualization of tissue structures.51 After intravenous injection of UCNPs-DI into the MCF-7 tumor-bearing nude mice which were pretreated with either UCNPs or UCNPs-DI, the PA images of tumor region were recorded at times of post-injection in 0, 1, 4, 8, 12 and 24 h, respectively. As depicted in Fig. 5b,d, only very slight PA signal increase was observed over time in the UCNPs cohort. In contrast, the PA signal in the tumor region of the UCNPs-DI pretreated cohort started to be detectable at 4 h post-injection, and the PA signal intensity gradually increased and reached the maximum at 12 h post-injection (Fig. 5c,d). The PA signal in the UCNPs-DI cohort at 12 h post-injection was determined as 10.8 times greater than that in the UCNPs controls (Fig. 5e). Moreover, the PA signal strength still maintained a relative high level after 24 h post-injection of UCNPs-DI, which further proves UCNPs-DI as an excellent PA diagnosing agent. Taken together, these results indicate that UCNPs-DI provide outstanding real-time PA imaging for definition of the tumor region and precise guidance for the subsequent laser irradiation.
2.5 Orthogonally regulated tumor targeting and therapy in vivo
We further evaluated the photoacoustic imaging-guided “on-demand” PDT efficacy in vivo by using the MCF-7 tumor-bearing mice as the animal model (Fig. 6a).52 The mice were randomly divided into six groups. The control group was injected with PBS; three groups received laser (PW/CW, 980 nm, 0.5 W/cm2 for 10 min) or UCNPs-DI i.v. injected alone; the other two groups received UCNPs-DI and then went through PW/CW corresponding irradiation (980 nm, 0.5 W/cm2 for 10 min). The changes in tumor volume and body weight were then monitored and recorded over the next 21 days. As shown in Fig. 6b, c, after injection of the nanoagent for 12 h and laser exposure to the pulsed laser, no obvious tumor growth inhibition effect was observed on during the whole treatment. By contrast, the tumor received the nanoagent and CW 980 nm laser treatment, the volume shrunk persistently, and the tumor growth was almost completely inhibited after treatment for 21 days, as also depicted in Fig. 6d. Figure 6e shows few changes in the body weight of the mice during feeding time, indicating that UCNPs-DI possesses high therapeutic efficiency and the systemic toxicity of the UCNPs-DI nanoagent with PW/CW irradiation was insignificant.
In order to further confirm the anti-tumor effect of UCNPs-DI, we used immunohistochemical staining for histological analysis to evaluate the pathological changes of the tumor tissue collected after the above treatment (Fig. 6f). Tumor sections revealed almost no cell necrosis and apoptosis in the nanoagent cohort with PW irradiation treatment, while the highest level of tumor cell damage was observed in the nanoagent cohort with CW irradiation treatment. In addition, we also performed hematoxylin and eosin (H&E) staining assays to study tumor cell death and organ damage after the phototherapy on the 21st day (Fig. 6g). The images of H&E staining demonstrated the tumor in the CW laser PDT group revealed significantly inhibited proliferation capability of the tumor cells, which is in agreement with the therapeutic results of the above-mentioned immunohistochemical staining in vivo. The photographs of the pathomorphological analysis of heart, liver, spleen, lung, and kidney are provided in Figure S17, which shows no obvious organ damage in all cases, indicating that the prepared UCNPs-DI nanocomposite have very good biocompatibility in vivo. Taken together, these results reveal that we can achieve the in vivo orthogonal safe PA imaging and highly effective phototherapy in the UCNPs-approved nanoagent through programmablly switching the on-duty ratio of the excitation laser with the same wavelength.