Design, preparation and characterization of Pd/H-TiO2 nanosonosensitizers
The Pd/H-TiO2 NSs were fabricated using a facile and mild hydrogenation maneuver according to the literature . In detail, the regular white TiO2 NSs were initially synthesized via a simple hydrothermal route. Next, just a slight amount of Pd was incorporated on the surface of TiO2 to form Pd/TiO2. Afterwards, H2 flow was fed without involvement of harsh terms including particularly high pressure and temperature and excitingly the yellow Pd/TiO2 transformed into black Pd/H-TiO2 (Fig. 1a and Additional file 1: Figure S2). Eventually, NH2-PEG2000 was adopted for surface modification of Pd/H-TiO2 to improve the dispersity and biocompatibility.
It could be seen from the transmission electron microscope (TEM) image that Pd/H-TiO2 NSs were of uniform rectangular nanostructure with the average diameter of 61.9 nm (Fig. 1b and Additional file 1: Figure S3). Although only a small quantity of Pd was introduced in the nanosystem, it was distinctly espied on TiO2 substrate (Fig. 1c), indicating that Pd could be tightly anchored on TiO2 matrix. From TEM images, both Pd/H-TiO2 and TiO2 were rectangular in shape (Fig. 1b and Additional file 1: Figure S4a), but from high-resolution transmission electron microscope (HRTEM) images, the superficial structure of Pd/H-TiO2 markedly differed from TiO2. Compare with the well-crystallized TiO2 (Additional file 1: Figure S4b), Pd/H-TiO2 had a typical amorphous shell (Fig. 1d) which was ascribed to the existence of oxygen defects. This could be also verified by scanning electron microscope (SEM) images, from which we could see the sharp edge of TiO2 NSs (Additional file 1: Figure S5) while the margin of Pd/H-TiO2 NSs was blunter (Additional file 1: Figure S6). The thickness of Pd/H-TiO2 analyzed by atomic force microscope (AFM) was about 1.3–1.5 nm (Fig. 1e, f). The energy dispersive spectrometer (EDS) in conjunction with elemental mapping clearly showed the coexistence of titanium (Ti), oxygen (O) and palladium (Pd) elements in Pd/H-TiO2 (Fig. 1g, h), and no Pd element in pristine TiO2 (Additional file 1: Figure S4c).
To shed light on the internal mechanism of the effect of H2 treatment on anatase TiO2, we performed density functional theory calculations with the Hubbard correction (DFT + U) on the electronic structures of the anatase TiO2(001) and TiO2(001)-2H (Additional file 1: Figure S1). The top, front and side views of these two models were shown in Fig. 2a, d, from which we could conclude that the adsorption of H atoms on the surface led to the deformation and reconstruction of the top two layers. Moreover, we plotted their 2D charge density maps at the same x coordinate position (Fig. 2c, f) to compare their density distribution difference. Obviously, the electrons rearranged because of the reconstructed top two layers, especially accumulated on the O atoms around H atoms. This phenomenon resulted in a new electronic band appearing at the middle of the bandgap of the TiO2(001)-2H compared with the original TiO2(001), which narrowed the bandgap from 2.56 eV to 1.66 eV and the long wave absorption was enhanced correspondingly (Fig. 2b, e). Simultaneously, we also calculated the total energy of H2 adsorption process on the TiO2(001) surface to verify the possibility of hydrogen atom adsorption. As shown in Fig. 2g, for the initial structure, there was a H2 molecule far away from the TiO2(001) surface. The H2 molecules were then moved toward the surface with a slight energy drop of 0.11 eV. When the H2 molecule was dissociated into two H atoms, one H was connected to the O atom, resulting in the deformation of the surface structure near the O, while the other H was connected to the adjacent Ti, and the energy drop of this process was 0.31 eV. When the second H atom desorbed from the Ti atom and then connected with the other nearest neighbor O atom, it would bring a higher energy drop (about 0.36 eV) and lead to more significant atomic reconstruction. These three energy drops demonstrated that it was energetically favorable of the reconstruction of the top two layers resulted from the adsorption of H2 molecules on the TiO2(001) surface.
Accordingly, the change of the electronic band structure did induce the differences of characterizations between Pd/H-TiO2 and TiO2. As suggested by X-ray diffraction (XRD) analysis, the diffraction peaks of Pd/H-TiO2 were in line with those of TiO2 but the peak heights dropped markedly (Fig. 2h), implying that the essential crystalline structure of Pd/H-TiO2 remained unchanged while the crystallinity decreased thanks to the formation of the amorphous layer. As indicated by X-ray photoelectron spectroscopy (XPS) analysis, the peak positions of Ti 2p (Additional file 1: Figure S7a, b), O 1 s (Additional file 1: Figure S7c, d) and C 1 s (Additional file 1: Figure S7e, f) in TiO2 and Pd/H-TiO2 were basically coincident, illustrating their similar binding energies with the environment. Additional file 1: Figure S8 disclosed the emergence of high valence state of Pd, which may be pertinent to the dissociation of H2. Specially, the valence band spectra uncovered the blue-shift of the valence band maximum of Pd/H-TiO2 (Fig. 2i), justifying the narrowed bandgap of Pd/H-TiO2. Besides, the narrowed bandgap of Pd/H-TiO2 was proved by the UV–vis-NIR absorbance spectra as well, from which the higher optical absorption of Pd/H-TiO2 over a wide wavelength range was readily observed (Fig. 2j). These results together revealed the process of TiO2 hydrogenation, that is, the process of white TiO2 turning into black Pd/H-TiO2. Specifically, the employed Pd induced the dissociation of H2 molecules on its surface, and then the dissociated hydrogen species spread to TiO2, reducing Ti4+ to Ti3+ accompanied by the generation of oxygen deficiencies. This process further resulted in the formation of a middle electronic band which finally narrowed the bandgap of TiO2 and augmented the visible-light absorption.
In vitro SDT, CDT and nanozyme efficiency of Pd/H-TiO2 nanosonosensitizers
Inspired by the existence of oxygen deficiencies and narrowed bandgap of Pd/H-TiO2 nanosonosensitizers, the SDT and CDT performances were then explored by evaluating the generation of ROS including 1O2 and •OH. For SDT performance, the 1O2 formed through the reaction of e− and oxygen molecules under US stimulation was initially detected using DPBF as the probe, based on the fact that the characteristic absorption peak of DPBF at around 420 nm in the UV–vis-NIR spectrum would decrease in the presence of 1O2. As expected, after irradiating the mixture of DPBF and Pd/H-TiO2, the absorbance intensity of DPBF apparently declined along with the extension of exposure time, and the descend range was wider in comparison with the mixture of DPBF and TiO2 at the same conditions (Fig. 3a, b). It could be seen from the results that Pd/H-TiO2 manifested enhanced efficiency of 1O2 production which was attributed to the disordered layer with abundant oxygen defects. Apart from 1O2, •OH was another member of ROS family, forming through the reaction of h+ and the surrounding water molecules under US excitation. The MB was chosen as the probe of •OH, for its degradation in the solution containing •OH would lead to the fall of the absorption peak at around 664 nm. Likewise, the •OH content in Pd/H-TiO2 solution was pronouncedly higher than that in TiO2 solution (Fig. 3c, d). The overt high yield of 1O2 and •OH guaranteed the eximious SDT performance of Pd/H-TiO2. On account of the formation of oxygen defects, it was presumed that there would be the coexistence of Ti3+ which could participate in Fenton-like reaction making contributions to •OH generation. Unsurprisingly, the downtrend of the absorption peak of MB was assuredly espied when Pd/H-TiO2 was blended with H2O2 at pH 5.5 (Fig. 3e), proving its CDT ability. And peculiarly, the downward tendency was more pronounced when US was imposed at the same conditions (Fig. 3f), suggesting the synchronous augment of SDT and CDT. The bilateral enhancement of SDT and CDT was further supported by ESR with TEMP and DMPO as the trapping agent of 1O2 and •OH, respectively. It was found that the maximum signal intensity appeared in Pd/H-TiO2 + H2O2 + US group no matter for 1O2 or •OH (Fig. 3g, h). The outcomes were consistent with that of UV–vis-NIR spectra to jointly confirm the reinforced SDT and CDT of Pd/H-TiO2.
As a type of precious metal, Pd has been comprehensively applied in biological areas considering its enzyme-mimetic activities [39,40,41]. Here, its catalase-like ability which meant the ability of catalyzing H2O2 into oxygen was tested by using a dissolved oxygen meter to monitor the change of oxygen concentration after adding H2O2 into Pd/H-TiO2 and TiO2 solutions under stirring. As shown in Fig. 3i, the oxygen concentration in Pd/H-TiO2 solution was constantly rising with time while there was no noticeable uptrend of oxygen concentration in TiO2 solution. Furthermore, CEUS was utilized to intuitively display oxygen production from H2O2, Pd/H-TiO2 and Pd/H-TiO2 + H2O2 solutions, for bubbles’ excellent backscattering capability could remarkably intensify the echo signal [42,43,44]. We observed the most obviously enhanced signal in Pd/H-TiO2 + H2O2 group from Additional file 1: Figure S9, which vividly substantiated the catalase-like capability of Pd/H-TiO2.
In vitro SDT and CDT effects against tumor cells
To reveal the antitumor potential of Pd/H-TiO2-PEG, the toxicity evaluation at the cellular level ensued. Certainly, affirmative intracellular uptake of Pd/H-TiO2-PEG was the precondition to ensure the effective killing to cancer cells. To acquire the evidence of endocytosis, cells were observed under the bio-TEM after coincubation without and with Pd/H-TiO2-PEG for 24 h. Encouragingly, pronounced clusters of Pd/H-TiO2-PEG (indicated by blue arrows) were easily found in the cytoplasm (Fig. 4a), affording a strong basis for Pd/H-TiO2-PEG to play its tumor killing role. Biocompatibility of nanoparticles was a vital indicator for future biological application. Here, the CCK-8 assay was used to assess the relative viabilities of human umbilical vein endothelial cells (HUVECs), 4T1 breast cancer cells and C6 glioma cells after cocultivation with increased doses (0, 12.5, 25, 50, 100, and 200 μg mL−1, based on Ti) of Pd/H-TiO2-PEG for 24 h. No matter for normal HUVECs, 4T1 or C6 cancer cells, Pd/H-TiO2-PEG alone showed no marked toxicity (Fig. 4b), reminding the innocuity of Pd/H-TiO2-PEG when lacking external incentives. However, when stimuli including H2O2, US and H2O2 + US were imposed after the coculture of C6 cells with different doses of Pd/H-TiO2-PEG, cell viabilities gradually declined with the elevated concentrations (Additional file 1: Figure S10) due to the generation of ROS through CDT and SDT. For further cytotoxicity investigation of SDT, CDT and SDT + CDT, cell viabilities were compared after experiencing diverse managements. As exhibited in Fig. 4c, cell viabilities significantly reduced in SDT group (group 7, Pd/H-TiO2-PEG + US) and CDT group (group 4, Pd/H-TiO2-PEG + H2O2), and the viability decline was even more in SDT + CDT group (group 8, Pd/H-TiO2-PEG + H2O2 + US), uncovering the synergetic enhancement of SDT and CDT.
ROS was the key medium for Pd/H-TiO2-PEG to assault cancer cells, therefore the intracellular ROS level was subsequently decided using the general probe DCFH-DA of which the hydrolysate in cells could be oxidized to fluorescent 2,7-dichlorofluorescein (DCF) by ROS. When cells were treated with Pd/H-TiO2-PEG + US, the strong green fluorescence was explicitly noticed, indicating a large deal of ROS production through SDT. The weaker fluorescence in Pd/H-TiO2-PEG + H2O2 group suggested lower ROS output through CDT than through SDT, but when US was combined, the fluorescence was conspicuously highlighted (Fig. 4d, e), certifying the boost of ROS generation by US and the collaborative effect of SDT and CDT. To present the cellular state after various disposals, the Calcein-AM and PI double staining was adopted to stain live and dead cells. The piles of dead cells (red) in Pd/H-TiO2-PEG + H2O2 + US group (Fig. 4f) visually reflected the effective cell killing effect of sono-chemodynamic therapy. The synergy of SDT and CDT was further confirmed by flow cytometry apoptosis analysis in the light of the Annexin V-FITC and PI costaining procedures (Fig. 4g).
Transcriptome alteration induced by combinatorial SDT and CDT
To get a deep insight into the intrinsic mechanism of Pd/H-TiO2-PEG mediated combinatorial SDT and CDT therapy in vitro, RNA sequencing was performed to analyze the gene expression in C6 cells after being treated (treatment group) and not being treated (control group) with Pd/H-TiO2-PEG + H2O2 + US. A total of 329 differentially expressed genes (DEGs) were ascertained, with 135 downregulated genes (blue) and 194 upregulated genes (red) in the cells of treatment group compared with the control group based on the settings of p-adjust < 0.05 and absolute fold-change cut off ≥ 1.5 (Fig. 5a). The clustering analysis clearly screened and separated the total upregulated and down regulated genes in cells of the two groups (Fig. 5b). From Fig. 5c, we could see significant differences in the expression of genes correlated with immune, apoptosis and tumor microenvironment (TME), indicating that Pd/H-TiO2-PEG mediated combinatorial SDT and CDT could function through regulating chemokines (CXCL1, CCL7, CXCL10, etc.), immune response (HIP1R, CSF3, CD28, etc.), epithelial-mesenchymal transition (CD44, ITGA5, SDC4, etc.), and cell adhesion (SERPINE1, TFRC, IBSP, etc.). Specially, owing to the creation of ROS during treatment, oxidative stress related genes (PLK3, NR4A3, KDM6B, etc.) involved in DNA repair were upregulated, suggesting the damage of ROS to C6 cells. In addition, the treatment could specifically induce the alteration of glioma related genes (ERBB2, NOTCH2, PDGFB, etc).
Gene ontology (GO) enrichment analysis distinctly explicated that DEGs were mainly concentrated in cell differentiation, intracellular signal transduction, epithelium development, cell proliferation and protein phosphorylation (Fig. 5d), all of which were indispensable biological processes for tumorigenesis and progression. The result of Kyoto encyclopedia of genes and genomes (KEGG) enrichment illuminated that DEGs were mostly implicated in MicroRNAs in cancer, cytokine-cytokine receptor interaction, TNF signaling pathway, IL-17 signaling pathway and MAPK signaling pathway (Fig. 5e). Generally, the enrichment of DEGs in TNF and IL-17 signaling pathways supported that Pd/H-TiO2-PEG + H2O2 + US treatment promoted tumor cell death mainly through the mechanism of regulating cell apoptosis and antitumor immune response. Specially, it could be inferred from the enrichment of MAPK signaling pathway that ROS triggered oxidative stress was a powerful driver of cell apoptosis. In addition, the HIF-1 signaling pathway was enriched, directly affirming the crucial role of oxygen in cancer therapy. The protein–protein interaction (PPI) network elucidated the interconnection between the proteins encoded by DEGs (Fig. 5f), laying the foundation for the in-depth mechanism research of the treatment.
In vivo biosafety, pharmacokinetics and biodistribution of Pd/H-TiO2-PEG
The biosafety of nanomedicines was a paramount prerequisite for in vivo application and future clinical translation. Therefore, the systemic toxicity of Pd/H-TiO2-PEG was assessed in priority. Briefly, after intravenous administration of different doses of Pd/H-TiO2-PEG on healthy mice, the body weight of each mouse was regularly recorded during 30 days, and blood indexes together with organ pathology were examined in the end. As shown in Additional file 1: Figure S11, body weights of the mice were constantly on the rise and there was no obvious discrepancy between all groups within the whole observation period. From the H&E staining images, we barely perceived any histological changes of major organs (heart, liver, spleen, lung and kidney) in the mice (Additional file 1: Figure S12). Also, the blood test results displayed that the haematological variables as well as the parameters of hepatic and renal functions were within normal limits and no relationship was presented between Pd/H-TiO2-PEG doses and the indexes (Additional file 1: Figure S13). The above results collectively illustrated the assured biosafety of Pd/H-TiO2-PEG, and this precondition laid the groundwork for the follow-up in vivo steps.
The pharmacokinetics and tissue distribution studies were conducive to disclosing the accumulative and metabolic processes of Pd/H-TiO2-PEG, being essential parts for guiding to establish an optimal treatment regimen. By detecting the Ti content of blood samplings collected at different time points post intravenous injection of Pd/H-TiO2-PEG, we got the blood circulation and eliminating rate curves of Pd/H-TiO2-PEG, which precisely told that Pd/H-TiO2-PEG had a relatively long blood half-life of 3.54 h and the eliminating rate descended from 0.39 μg mL−1 per h to 0.02 μg mL−1 per h at 1.36 h (Additional file 1: Figure S14). Thereupon, the biodistribution of Pd/H-TiO2-PEG in major organs and the tumor site was investigated on C6 tumor-bearing mice. It turned out that Pd/H-TiO2-PEG could overtly enrich in the tumor at 24 h post injection, contributing to deciding the rational interval between the medication and US irradiation (Additional file 1: Figure S15).
Pd/H-TiO2-PEG NSs mediated in vivo tumor inhibition
The satisfying in vitro therapeutic efficacy and comprehensive in vivo pre-work massively propelled the exploration of the in vivo antitumor action of Pd/H-TiO2-PEG. The BALB/c nude mice bearing subcutaneous C6 tumors were randomly divided into five groups (n = 5 in each group): (1) Control (PBS), (2) US, (3) Pd/H-TiO2-PEG, (4) TiO2 + US and (5) Pd/H-TiO2-PEG + US. The dosage of TiO2 and Pd/H-TiO2-PEG was the same as 100 μL (10 mg kg−1, i.v. injection) and the US settings were 1.5 W cm−2, 1.0 MHz, 50% duty cycle and 3 min. The mice in group 1, 3, 4 and 5 firstly underwent intravenous administration of varied reagents and then the mice in all groups were treated with or without US at 24 h post injection. The same protocols were executed on the second day and the whole observation period was 14 days (Fig. 6a), during which the body weight and tumor dimensions of each mouse were monitored every other day. The growth curves of individual tumor (Fig. 6b) and each group of tumors (Fig. 6c) manifested affirmative tumor restraint in Pd/H-TiO2-PEG group (CDT), TiO2 + US group (SDT) and especially in Pd/H-TiO2-PEG + US group (CDT + SDT), with the tumor inhibition rate of 34%, 61% and 90%, respectively (Fig. 6d), while body weights of the mice steadily increased with time (Additional file 1: Figure S16). The results corroborated the indisputable antitumor efficacy of bilaterally enhanced SDT and CDT with neglectable side effects. The CEUS images of the tumor before and after Pd/H-TiO2-PEG + US treatment revealed evident reduction of the internal blood supply (Fig. 6e), confirming the destruction of tumor vasculature. Furthermore, at the end of the observation period, the lightest weight and minimum size of the tumors in Pd/H-TiO2-PEG + US group also offered convincing proofs for the therapeutic effect of synergistic SDT and CDT (Fig. 6f-h).
Ultimately, tumor tissues in all groups underwent H&E, TUNEL, Ki-67, CD31 and HIF-1α staining for exhaustive histological analyses, and the pathological changes could be straightly identified from Fig. 6i and Additional file 1: Figure S17. From H&E and TUNEL staining images we could easily observe the chromatin condensation, nuclear disintegration and strongest green fluorescence signal in Pd/H-TiO2-PEG + US group, attesting that Pd/H-TiO2-PEG + US disposal brought about severe cell apoptosis. The significantly decreased red fluorescence signal of Ki-67 and CD31 suggested the inhibition of cell proliferation and tumor angiogenesis. The eximious oxygen supply ability of Pd/H-TiO2 has been verified in vitro. Here, the in vivo hypoxia remission capability was testified by detecting HIF-1α, a key factor that highly expressed in hypoxia tissues to regulate cells’ adaptive responses to low oxygen tension. From Fig. 6i, we could see the most conspicuous red fluorescence signal of HIF-1α in control group, reflecting severe tissue hypoxia. In comparison, the hypoxic condition was significantly alleviated in Pd/H-TiO2-PEG and Pd/H-TiO2-PEG + US groups thanks to the catalytic action of Pd component. Noteworthily, in the TiO2 + US group, the expression of HIF-1α was also decreased compared with the control group, which had nothing to do with Pd. For this, it was deemed that the hypoxia amelioration was credited to the reduced oxygen consumption following tumor shrinkage caused by TiO2 + US induced SDT effect. Except for the histological proof, the enhanced CEUS signal after intratumorally injection of Pd/H-TiO2-PEG also intuitively testified its intratumoral oxygen generation property (Additional file 1: Figure S18), which was crucial assistance for the reinforcement of SDT and CDT. Besides, no visible pathological damage of major organs in all groups was perceived through H&E staining (Additional file 1: Figure S19), implying the indubitable therapeutic safety. On the whole, Pd/H-TiO2-PEG NSs could be used as innocuous and reliable sonosensitizers with integrated SDT, CDT and oxygen production effects to achieve a satisfactory therapeutic efficiency for cancers.