To obtain Pt Cs with uniform and closed-packed deposition of Pt NPs, we adopted laser irradiation to generate GBs-enriched Mn3O4 NPs for subsequent galvanic replacement reaction. Figure 1(a) schematically illustrates the formation processes from completely crystallized raw Mn3O4 (R-Mn3O4) to laser-treated Mn3O4 (L-Mn3O4) with enriched GBs and finally to the desired Pt Cs. Accordingly, the corresponding sample changed from yellow suspension to orange colloidal solution, and finally to black colloidal solution (Figure S1). For metal oxide NPs, absorbing laser with photon energy (hν) larger than the bandgap (Eg) can immediately promote strong photoionization, resulting in structure fragmentation into educts. These educts can accumulate and regrow as novel NPs with incomplete crystal structure55. Herein, we found that the R-Mn3O4 NPs exhibited strong absorption of light from 200 to 450 nm, and the Eg of R-Mn3O4 NPs was calculated as 2.54 eV56 (see Figure S2 for details), so a pulsed laser beam with a wavelength of 355 nm (photon energy of 3.49 eV) was selected as the irradiation source. The transmission electron microscopy (TEM) images in Figure S3 show the nice crystallinity for R-Mn3O4 NPs with a size of 50 nm. Figures 1(b) and 1(f) reveal that L-Mn3O4 NPs have a similar size and composition compared to R-Mn3O4 NPs, but GBs are generated on the particle surface after laser irradiation in Figures 1(c)–1(d). In our experiments, with laser induced photoionization, R-Mn3O4 NPs were broken into metastable educts in an aqueous environment (Figure S4), and then the educts underwent ripening and regrowth processes to transform into GBs-enriched L-Mn3O4 NPs57. By analyzing the lattice structure for L-Mn3O4 NPs, we found that the {200} and {004} planes were the two major exposed planes, and GBs emerged between the two planes (see Figure S5 for details). Furthermore, the comparison between oxygen species for Mn3O4 before and after laser irradiation indicates the existence of abundant GBs (see Figure S6 for details).
To further study the impact of GBs in the deposition of Pt NPs, density function theory (DFT) calculations of absorption energy for PtCl42− groups on different planes and the GBs were compared. According to the shape of tetragonal R-Mn3O4 NPs, the {200} and {004} planes were the two major exposed planes (Figure S7). In Figure 1(e), the adsorption energy values of PtCl42− groups on the {200} and {004} planes on the Mn3O4 surface were recalculated to be −9.90 eV and −3.88 eV, respectively. This means that PtCl42− groups tend to be adsorbed on the {200} planes. Moreover, the site for Pt deposition is decided by the Mn2+ ions on the surface of Mn3O4 NPs in the galvanic replacement reaction44. The number of Mn2+ ions accounts for two-fifths on the {200} plane and one-half on the {004} plane. When PtCl42− groups react with limited Mn2+ ions on the {200} plane, the formed Pt NPs easily deposit on some local regions near Mn2+ ions and have uncontrollable size. However, laser-induced GBs indicate that the miniaturization of grains can promote the random distribution of the {200} and {004} planes to support even deposition of ultrasmall Pt NPs48. In addition, GBs are conducive to the galvanic replacement reaction, because they are beneficial for exposing more active sites58–60. GBs also favorably alter metal-oxygen bonds and local electron density by inducing surface compressive strains, thereby indirectly tuning the electron configuration and the adsorption ability to surface adsorb reactants51–53. The calculated adsorption energy of PtCl42− groups on GB was to be −4.51 eV. Therefore, L-Mn3O4 with abundant GBs can improve the galvanic replacement reaction with PtCl42− groups, which is helpful for obtaining Pt Cs with high density and uniform Pt NPs loading.
As expected, Pt Cs possess a porous inner structure and an outer shell of closed-packed ultrasmall Pt NPs. Figure 1(f) shows that Pt Cs look like spherical NPs with an average size of 55 ± 6.6 nm. The energy-dispersive X-ray spectroscopy (EDX) mapping analysis in Figure 1(g) confirms the uniform distribution of Mn, Pt, and O, and the X-ray photoelectron spectroscopy (XPS) spectra in Figure S8 prove that the Pt and Mn in Pt Cs are Pt0 and a mixture of Mn+2 and Mn+3 species, indicating that the Pt and Mn exist as metal and oxide, respectively. Based on inductive coupled plasma emission spectrometer (ICP) data, the molar ratio Pt/Mn can reach as high as 2.23, but the value for R-Pt Cs synthesized from R-Mn3O4 under the same conditions only reaches 1.06 (Figure S9(a)). This indicates that the deposition rate of the galvanic replacement reaction on R-Mn3O4 NPs was much lower than that on L-Mn3O4 NPs. This phenomenon is also consistent with the Brunauer–Emmet–Teller (BET) method, which shows that Pt Cs have a larger porous inner structure (specific surface area is 101.7 m2·g-1, see Figure 1(i)) compared to commonly etched R-Pt Cs (specific surface area is 65.15 m2·g-1, see Figure S9(b)). The most part of Mn3O4 in Pt Cs was self-sacrificed during the galvanic replacement reaction, which is why the characteristic peaks of Mn3O4 almost disappeared in the X-ray diffraction (XRD) patterns (Figure 1(h)). As shown in Figure 1(j), the deposited Pt NPs on the Pt Cs surface had an average size of 2.17 ± 0.03 nm, and the measured interplanar spacing was about 0.225 nm, which corresponded to the (111) plane of cubic Pt NPs, consistent with the broadened Pt diffraction peaks (JCPDS No. 96-101-1104) in the XRD pattern of the Pt Cs. Additionally, the ultrasmall Pt NPs were closely packed with a tiny interparticle distance of 0.29 ± 0.04 nm. As shown in Figures S9(c)–S9(d), the size of deposited Pt NPs on R-Pt Cs and their interparticle distance were 2.63 ± 0.08 nm and 1.29 ± 0.1 nm, respectively. Both were larger than the corresponding values for Pt Cs. Ultrasmall Pt NPs (≤ 5 nm) are extremely helpful for subsequent catalytic reactions. More importantly, the generation of hot electrons in NPs is a quantum effect that is strongly size-dependent61–62. In addition, the uniform and dense arrangement of Pt NPs is conducive to near-field coupling between adjacent Pt NPs, thereby expanding the broadening of light absorption for the Pt Cs25, 63–65. The porous structure is also conducive to the continuous reflection of light in Pt Cs, thus greatly enhancing the light absorption efficiency of Pt Cs66–68. All of the preceding analysis lays the foundation for the generation of photo-excited hot electrons.
The generation of photo-excited hot electrons in noble metal systems needs to meet two preconditions: first, the metal can absorb photons with a specific wavelength; second, the photon energy is high enough to excite the d-band electrons to cross the Fermi level6, 69. As shown in Figure 2(a), Pt Cs indeed expanded the light absorption from the visible band to the NIR band for ultrasmall Pt NPs (Figure S10) as expected. Considering the NIR band, Pt Cs have a higher absorption capacity than R-Pt Cs (Figure S15(g)). The electronic density of the state distribution of Pt, determined by DFT calculations and presented in Figure 2(b), showed a wide d-band distribution (−6.4 eV to 0.41 eV, DOS > 0.5) and a high peak of d-band reaching the Fermi energy. This confirms a very high density of electron states at the Fermi energy, and ensures that the energy of NIR photons (808 nm = 1.53 eV) can easily excite the electrons located at the part of the d-band below the Fermi energy for interband (d-s) transition or intraband (d-d) transition. The long-lived hot electrons excited by NIR photons can participate in the catalytic reaction in two ways6: one is the hot electrons with higher kinetic energy and higher momentum than other electrons at the Fermi surface may transfer energy to certain chemical bonds of molecules to break these bonds, such as in the decomposition of H2O2 into O2 and H2O; the other is the hot electrons can inject directly into unoccupied orbits of target molecules and lower the potential of the molecules required by chemical reactions, such as in the conversion of O2 to O2•−. In our experiments, Pt Cs have been proved as CAT and OXD mimics in catalyzing the decomposition of H2O2 and O2, respectively (see Figures S11–S12 for details). Pt Cs also exhibit the self-cascade catalytic activities to produce O2•−- from H2O2 in a simulated TME70–71. Therefore, the changes in the enzyme-like activity for Pt Cs can be used to investigate the effect of possible hot electrons on the enzyme-like activity.
We prepared four groups of Pt Cs samples to investigate their enzyme-like activities corresponding to different Pt/Mn ratios (see Table 1 and Figures S13–S14 for details). When the molar ratio of Pt/Mn elements increased to 2.2, the relative mass activity of the Pt in Pt Cs reached a peak, and the impact of the residual Mn3O4 content could be eliminated due to tight surrounding of ultrasmall Pt NPs. The enzyme-like catalytic activities of Pt Cs and other controls with and without NIR light irradiation (808 nm laser as light source) were compared under room temperature. For CAT mimic, OXD mimic, and self-cascade catalytic reactions, the activity of Pt Cs under NIR light irradiation at room temperature increased by 109.3%, 53.1%, and 41.8%, compared to the same situations without NIR light irradiation. The corresponding activity enhancement values were 55.7%, 38.3%, and 18.6% for R-Pt Cs and 37.8%, 10.6%, and 11.9% for Pt NPs, respectively. These results indicated that the activity of Pt Cs significantly increased under NIR light irradiation, and the increased values were higher than those of R-Pt Cs and Pt NPs. In noble metals, the absorption of photons can lead to the generation of hot charges or the generation of heat by electron-phonon coupling72. To evaluate the heat conversion, we tested the photothermal effect of Pt Cs and R-Pt Cs under NIR light (see details in Figure S15). The photothermal conversion efficiencies of Pt Cs and R-Pt Cs were calculated to be 23.9% and 18.9%, respectively.
To further explore the effect of heat induced by NIR light on the enzyme-like activity, we simulated the photothermal effect through heating in a water bath73. Figures 2(c)–2(e) demonstrate that the pure thermal effect limited the promotion of the activity of these samples. Taking CAT mimic as an example, the activity enhancement of Pt Cs affected by thermal effect was 76.7%, which was 32.6% lower than the contribution of NIR light. Moreover, the difference between the contribution of NIR light and thermal effect on the enzyme-like activity was 11.8% for R-Pt Cs and 3.5% for Pt NPs. These significant differences between the contribution of NIR light and thermal effect can be attributed to the hot electrons excited by NIR photons. For OXD mimic and self-cascade catalysis, hot electrons increased the catalytic activity of Pt Cs by 25.2% and 15.3%, respectively, that of R-Pt Cs by 13.9% and 5.6%, respectively, and that of Pt NPs by 3.4% and 2.5%, respectively. Compared with Pt Cs, the activity enhancement induced by hot electrons for R-Pt Cs was rather weak, while the increase in the activity of Pt NPs was almost negligible. Pt Cs can support a high efficiency for the generation of hot electrons, thereby facilitating the improvement of enzyme-like catalytic activity.
We also carried out the reverse verification experiments under an ice bath condition (Figures 2(f)–2(h)) to exclude the contribution of the heat converted from the NIR light. Compared with the ice bath groups in the dark, the ice bath groups irradiated by the 808 nm laser successfully increased the CAT mimic activity of Pt Cs by 73.3%, increased the OXD mimic activity by 38.8%, and increased the self-cascade catalytic activity by 30.3%. The enhancing trend of enzyme-like activity between different controls is consistent with the results above, which further verifies the contribution of hot electrons to enzyme-like activity. Visible light was also used as a light source to detect the enzyme-like activity of Pt Cs with a negligible photothermal effect (Figures S16(a)–S16(c)). The activity enhancement of Pt Cs under visible light further verifies the positive effect of hot electrons. More details associated with these experiments are presented in Figure S16.
To further confirm the conversion from NIR photons to hot electrons, Pt Cs, R-Pt Cs, Pt NPs, and Mn3O4 NPs were tested in view of the photocurrent in Figure 2(i). The photocurrent was induced by hot electrons immediately upon illumination and reached its maximum magnitude within several milliseconds. It usually takes longer (on the order of seconds) to reach the thermal diffusion equilibrium for the heat-induced temperature increase of the surroundings74–75. Accordingly, a photoelectrochemical method with a millisecond time resolution would be able to disentangle the charge carrier effect from the photothermal effect. Figure S17 shows the Iphoto-t curve of a Pt Cs electrode in a single current response process. The current reached the peak immediately upon 808 nm laser irradiation and then returned to the initial state immediately when the laser was turned off. The response and recovery times of the photocurrent were both at the millisecond level, which is consistent with the theoretical response time of the photoelectric effect caused by hot electrons. Figure 2(i) compares the photocurrent signals of Pt Cs with other controls. The photocurrent density of Pt Cs was 1.8 µA·cm-1 under 808 nm laser irradiation, almost 3.6 times that of R-Pt Cs (0.5 µA·cm-1). The photocurrent signals of Pt NPs and Mn3O4 NPs were so weak that they could be ignored. This further indicates that Pt Cs can generate a large number of hot electrons under NIR photon-excitation.
Furthermore, enhancement of the electric field (|E|/|E0|) induced by NIR photon-excitation was calculated using the finite difference time-domain (FDTD) method, where |E| and |E0| denote the photo-enhanced and initial incident electric fields, respectively. Hot electron injection will lead to significant enhancement of the electric field, so the stronger the electric field in Pt Cs, the greater the number of hot electrons generated51, 76. Figures 2(j) and 2(k) show the illustration of the constructed model of Pt Cs and the calculated spatial distribution of the photo-induced electric field enhancement. The highest local electric field enhancement (|Emax|/|E0|) of Pt Cs was 6.88, which was far higher than 0.914 of R-Pt Cs and 0.596 of pure Pt NPs (shown in Figure S18). The location of highest |E|/|E0| was concentrated between the closed-packed Pt NPs on the surface of the Pt Cs, which was caused by the near-field coupling effect between adjacent ultrasmall Pt NPs63–65. Therefore, NIR photon-excited hot electrons play an important role in promoting enzyme-like catalytic activity.
Figure 3 displays the investigated enzyme-like catalytic kinetics for Pt Cs with and without NIR photon-excitation. First, the CAT mimic activity of Pt Cs was modulated by a laser with power density in the range of 0–3 W·cm-2. Figure S19 proves that an 808 nm laser and limited heat cannot lead to significant self-decomposition of H2O2 in systems. As shown in Figure 3(a), as the laser power density increased, so did the amount of O2 produced by the catalytic decomposition of a specific amount of H2O2 in solution (100 µM) in 10 min. Compared with the system without laser irradiation, the concentration of generated O2 under laser irradiation (2 W·cm-2) increased to 203.0%, and even to 215.8% when the laser power density rose to 3 W·cm-2. Figure 3(b) monitors the change in H2O2 concentration according to its characteristic absorption peak at 240 nm77. The concentration of H2O2 decreased significantly with the increase in laser power density. Under 3 W·cm-2 laser irradiation, the absorption peak of H2O2 almost disappeared within 10 min, indicating that H2O2 was completely decomposed. Comparatively, the concentration of H2O2 in the system without Pt Cs had no significant decreases under laser irradiation with different power densities. The impact from NIR light irradiation to the OXD substrate o-phenylenediamine (OPD) in solution can be neglected (Figure S20). Figure 3(e) shows that, in the range of 0–2 W·cm-2, the characteristic absorbance of the oxidized OPD gradually increased with the increase in laser power density in 10 min. However, the absorption intensity of the oxidized OPD dropped when the 808 nm laser power density exceeded 2 W·cm-2. This means that the OXD mimic activity of Pt Cs increased first and then decreased with the increase in laser power density, which may be induced by temperature-dependence of nanozymes. Additionally, the electron spin resonance spectroscopy (ESR) spectra in Figure 3(f) compare the O2•− signals under different conditions77. The peak intensity of O2•− under NIR light irradiation was about 4.1 times higher than in dark conditions, indicating that NIR photon-excitation can indeed promote the generation of O2•−. The steady-state catalytic kinetics for Pt Cs were analyzed by fitting the Michaelis–Menten curves and Lineweaver–Burk plots (Figures 3(c)–3(d), Figures 3(g)–3(h)). The maximum velocity (Vmax) and Michaelis–Menten constant (Km) for CAT mimic behaviors of Pt Cs under NIR light irradiation were calculated as 7.14 × 10-4 M·s-1 and 0.35 mM, respectively78. For the control group without NIR light irradiation, the corresponding Vmax and Km values were 5.46 × 10-4 M·s-1 and 0.62 mM. In comparation, Vmax increased by 30.8%, and Km decreased by 43.5% under NIR light irradiation. Similarly, the Vmax and Km for OXD mimic behaviors before and after NIR light irradiation were 16.74 × 10-8 M·s-1, 0.056 mM and 21.77 × 10-8 M·s-1, 0.054 mM, respectively. The lower Km indicates a stronger affinity to the nanozyme substrates, and the larger Vmax indicates a faster speed in catalytic reactions. Thus, NIR light irradiation can obviously improve the affinity of H2O2 for Pt Cs and the chemical transformation of O2 on the surface of Pt Cs.
The self-cascade catalytic activity of Pt Cs in simulated TME was tested. As shown in Figure 3(i), the production of ROS increased with the increase in laser power density and reached its peak at 2 W·cm-2. Figure S21 compares the generated ROS levels in the OXD mimic catalysis test and the self-cascade catalysis test. The self-cascade catalytic activity of Pt Cs significantly promoted the generation of ROS in the reaction system. The ESR data in Figure 3(j) confirmed that the addition of H2O2 alone could not change the concentration of O2•− in solution, and neither the control group of Mn3O4 NPs nor Pt NPs could significantly promote the transformation from H2O2 to O2•− in solution. When Pt Cs emerged in the tested solution, the signal of O2•− was apparently enhanced in the ESR spectrum. NIR light irradiation further improved the peak intensity for O2•− by 2.9 times. Therefore, NIR light irradiation was confirmed to greatly enhance the self-cascade catalytic ability of Pt Cs to generate sufficient ROS to destroy tumor cells (Figure 3(k)).
Before the cell experiment, we used polyethylene glycol (PEG) to biomodify Pt Cs and verified the stability of PEG/Pt Cs in phosphate-buffered saline (PBS) buffer (Figure S22). Compared with normal cells, tumor cells produce excessive amounts of H2O2 (up to 0.5 nmol per 104 cells per h)79. H2O2 in TME can be converted into O2 by CAT mimic and then catalyzed by OXD mimic to produce numerous highly toxic ROS, which can effectively kill tumor cells. In this experiment, DCFH-DA was used react with ROS to generate 2,7-dichlorofluorescein (DCF) with green fluorescence, which enabled the detection of ROS in cells. As shown in Figure 4(a), negligible green fluorescence was observed in the control, 808 nm laser, and Mn3O4 groups. The Pt group showed fluorescence, and the fluorescence of 4T1 cells treated with PEG/Pt Cs was more obvious. Further, compared with other experimental groups, the PEG/Pt Cs + laser group showed the most obvious green fluorescence, indicating that the cells co-incubated with PEG/Pt Cs can produce a large amount of ROS under NIR light irradiation. The flow cytometry in Figure 4(b) quantifies the ROS of the above experimental groups. The ROS production rate of the PEG/Pt Cs + laser group (84.8%) outclassed that of the PEG/Pt Cs group (40.2%), Pt group (26.8%), and other control groups.
The effects of PEG/Pt Cs on the cell viabilities of normal cells and tumor cells were compared. As shown in Figure 4(c), the increase in PEG/Pt Cs concentration can decrease the tumor cell viabilities. Regarding the PEG/Pt Cs + laser group, its half maximal inhibitory concentration (IC50) was 83 µg·mL-1, which was far lower than that of the experimental group without NIR light irradiation (180 µg·mL-1). This further indicates that NIR light promotes the catalytic activity of Pt Cs and produces more ROS, leading to increased cytotoxicity. However, for normal cells, due to the low content of H2O2, the PEG/Pt Cs + laser group with different concentrations showed good safety and low toxicity in normal cells (293T and hl7702) (Figure 4(d)). Hence, the PEG/Pt Cs + laser group had good biosafety and low toxicity. To observe the living and dead cells, cells were co-stained with Calcian-AM and PI after treatment with different experimental groups. As shown in Figure 4(e), the PEG/Pt Cs + laser group exhibited more dead cells than the PEG/Pt group, pointing to the better antitumor activity of the PEG/Pt Cs + laser group.
As displayed in Figure 5(a), the therapeutic evaluations of PEG/Pt Cs were examined on 4T1 tumor inoculated mice. To design a more reasonable treatment scheme in vivo, we tracked the enrichment of Pt Cs at the tumor site through different imaging methods. Our previous study demonstrated that a payload of PtOx on porous Mn3O4 shows enhanced longitudinal (T1) contrast ability for magnetic resonance imaging (MRI)80. Figure S23(a) shows that the longitudinal relaxivity (r1) of PEG/Pt Cs is 3.5 mM-1·s-1, and the transverse relaxivity (r2) is 7.4 mM-1·s-1, resulting in the r2/r1 value of 2.11, which is suitable for T1 MRI81. Figure 5(b) shows T1 MRI of the tumor site at different times after intravenous injection of PEG/Pt Cs in mice. There was no obvious signal at the tumor site before injection, whereas the tumor began to show enhanced contrast within 0.5 h after PEG/Pt Cs injection. T1 contrast continued to increase until it reached the highest point at 6 h after injection. This may have been due to the passive accumulation of PEG/Pt Cs at the tumor site resulting from the enhanced permeability and retention effect. A quantitative analysis of the signal-to-noise ratio (SNR) in the tumor site was performed to evaluate the potential of PEG/Pt Cs as a contrast agent. Figure S23(b) shows that the SNR value increased from 52.7 (before injection) to 65.8 (6 h post injection) and then gradually decreased to 46.9 (24 h post injection). In addition, according to the MRI tracking data above, we used infrared thermography to observe the temperature change at the tumor site in mice after 6 h post injection. Figure 5(c) shows that under 808 nm laser irradiation, the temperature of the tumor site of mice injected with PEG/Pt Cs gradually increased with the increase in irradiation time, reaching 48.0℃ after 10 min of laser irradiation. In contrast, the tumor site of mice injected with PBS buffer was only 38.3℃ after 10 min of laser irradiation (Figure 5(d)). Moreover, infrared imaging of the dissected mice organs showed that the tumor signal was significantly higher than that of other organs (Figure S24). Our findings proved that PEG/Pt Cs accumulated at the tumor site 6 h after injection.
Due to differences in pH value and H2O2 content between tumor tissues and normal organs, cancer cells are more vulnerable to the impact of exogenous ROS compared with normal cells82. PEG/Pt Cs can generate a large number of ROS through efficient self-cascade catalysis under NIR photon-excitation, which has the potential to specifically inhibit tumor growth. As shown in Figure 5(e), during the 21-day treatment period, the average body weight of each group did not change significantly, indicating that the adverse reactions caused by PEG/Pt Cs injection during the anti-cancer treatment can be ignored. Compared with the other five groups, the PEG/Pt Cs + laser group had the best tumor treatment effect and significantly inhibited the tumor growth of mice (Figure 5(f)). The average tumor volume of the PEG/Pt Cs + laser group was 70.42 mm3 at 21 days after the first dose, much smaller than that of the PBS group (1200.95 mm3), PBS + laser group (1204.78 mm3), Mn3O4 group (921.80 mm3), Pt group (437.79 mm3), and PEG/Pt Cs group (204.83 mm3) (Figure 5(g)). Furthermore, Figure 5(h) indicates that PEG/Pt Cs had outstanding therapeutic outcomes on tumors under NIR light. Hematoxylin and eosin (H&E), Ki67, and TUNEL staining assays of tumor tissues further proved that the tumor cells underwent severe apoptosis after 21 days of treatment (Figure 5(i)). Notably, the blood biochemical and haematological analysis showed that there were no comparable differences between the PEG/Pt Cs-treated group and the control group (Figure S25), demonstrating no significant infection and inflammation were arose during the whole evaluation period. Moreover, H&E analysis was performed on the main organs (Figure S26). As with the control group, the mice treated with the PEG/Pt Cs exhibited no obvious damage to the heart, liver, spleen, lung, or kidney, indicating that PEG/Pt Cs had no obvious side effects on the physiological function or health status of mice during the treatment. The above data show that the process of antitumor inhibition by PEG/Pt Cs had low toxicity to organisms and good biosafety.