Simultaneous Inhibition of Heat Shock Response and Autophagy with Bimetallic Mesoporous Nanoparticles to Enhance Mild‐Temperature Photothermal Therapy

Mild‐temperature photothermal therapy (MPTT) is a promising tumor therapeutic modality because it can avoid the damage of normal tissues near the tumor caused by excessive heat. However, its therapeutic effect is severely impaired because tumor cells can develop heat resistance and self‐repair by activating heat shock response and cell autophagy. Herein, a tannic acid–iron ion metal organic framework‐coated, chloroquine (CQ)‐loaded mesoporous PdPt nanosystem (TF‐CQ@mPdPt) is developed to enhance MPTT by simultaneous suppression of heat shock response and autophagy. TF‐CQ@mPdPt exhibits good peroxidase (POD)‐mimic activity and photothermal performance. As a result, the reactive oxygen species generated by POD‐mediated decomposition of endogenous hydrogen peroxide damage mitochondria, leading to limitation of adenosine triphosphate supply, which suppresses the upregulation of heat shock proteins of tumor cells during MPTT, making tumor cell more sensitive to heat stress. Concurrently, CQ released from TF‐CQ@mPdPt during MPTT inhibits cell autophagy, thereby interrupting the self‐repair pathway of tumor cells. Consequently, TF‐CQ@mPdPt‐mediated MPTT significantly enhances its therapeutic effect, effectively inhibiting tumor progression in 4T1 tumor‐bearing mice. This study presents a novel strategy to enhance MPTT by simultaneously suppressing heat shock response and autophagy.


Introduction
As one of the most promising strategies, photothermal therapy (PTT) suppresses tumors by local heat when photothermal agents are irradiated by laser. [1,2]To effectively kill tumor cells, the local temperature of PTT is expected to be as high as possible, which usually damages the surrounding normal tissues during treatment, and might enhance invasion and metastasis of the residual tumor cells. [3]To overcome the defects of PTT, mild-temperature photothermal therapy (MPTT) is proposed, but lots of studies have suggested that MPTT alone was not promising due to the mild heat resistance of tumor cells. [4]hen heat stress is not fatal, tumor cells usually develop tolerance to mild heat for self-repairing. [5]During MPTT, heat stress induces tumor cells to express heat shock proteins (HSPs) to trigger heat shock response, relieving cell damage caused by local mild heat. [6]Take heat shock protein 70 (HSP70) as an example, its expression is at a low level in the cytoplasm under normal physiological conditions and is involved in protein de novo folding.When heat stress occurs, HSP70 expression increases significantly to save cells from heat damage by preventing the aggregation of unfolding proteins and refolding the aggregated proteins. [7]Studies have also proved that MPTT could enhance cell autophagy to reduce heat damage. [8]Autophagy is a cellular dynamic degradation pathway that recycles damaged cellular entities and proteins in response to various stresses. [9]When under stress (hunger, oxidative, heat, etc.), bilayer structures are generated in the cytoplasm to form autophagosomes, which wrap organelles and nonspecific proteins.Then, autophagosomes fused with lysosomes to form autolysosomes to degrade the wrapped contents. [10]Inhibition of autophagy by drugs or knockout of autophagy-related genes can interrupt the self-repair pathway of tumor cells, thus improving the efficacy of cancer therapy. [11]herefore, heat shock response and autophagy are responsible for the "mild heat resistance" of tumor cells to MPTT, and inhibiting heat shock response and modulating autophagy could be a potential strategy for improving MPTT.Nanoparticles (NPs) based on platinum (Pt) and palladium (Pd) usually have wide near-infrared (NIR) absorption and good photothermal conversion efficiency. [12]Meanwhile, they also have peroxidase (POD)-mimic activity, which can catalyze oxidation of tumor endogenous hydrogen peroxide (H 2 O 2 ) into hydroxyl free radicals (•OH). [13]Notably, recent studies have revealed that intracellular accumulation of •OH would oxidize mitochondria, resulting in mitochondrial damage and thus reduction in adenosine triphosphate (ATP) production. [14]The energy deficiency inhibits heat shock response as the cellular ATP level decreases. [15]imultaneously, excessive •OH can also react with proteins to form cross-links to destruct the structure and function of HSPs. [16]n addition, chloroquine (CQ), a widely used drug for treatment of malaria and autoimmune disease, is also an autophagy inhibitor for clinical therapies. [17]Therefore, we infer that coordination of POD-mimic enzyme property of Pd-Pt NPs with CQ could achieve heat shock response suppression and autophagy inhibition, enhancing the antitumor effect of MPTT.
Therefore, in this study, we developed a CQ-loaded mesoporous PdPt (mPdPt) nanosystem and coated it with a pH-responsive tannic acid-iron ion (TA-Fe (III)) metal organic framework (designated as TF-CQ@mPdPt).TF-CQ@mPdPt has good photothermal conversion performance and POD-mimic activity.Upon accumulation in the tumor, TF-CQ@mPdPt could achieve MPTT upon near-infrared (NIR) laser irradiation, produce cytotoxic •OH to suppress the expression of HSP70, and simultaneously release CQ to effectively inhibit autophagy, leading to the enhanced antitumor effect of MPTT (Scheme 1).

Results and Discussions
2.1.Synthesis and Characterizations of TF-CQ@mPdPt TF-CQ@mPdPt nanosystem was fabricated by first synthesis of mPdPt NPs and then sequentially loading CQ and coating with a TA-Fe (III) metal organic framework (MOF) on the surface (Scheme 1).mPdPt NPs were synthesized by a typical surfactantdirecting synthesis method, by which Pd species were first reduced by ascorbic acid (AA), forming Pd-rich metal seeds, and subsequently the self-assembled pluronic F127 surfactant micelles capped the Pd seeds, serving as templates for Pt deposition.The size of mPdPt NPs could be adjusted by tuning the acidity of the reaction mixture solution through adding different amount of HCl (Figure S1, Supporting Information). [18]When 15 μL of HCl was added, the average diameter of mPdPt NPs was about 100 nm with a narrow size distribution (Figure 1a and S2, Supporting Information).mPdPt NPs had a well-ordered mesoporous structure, which was piled up by ultrasmall NPs (Figure 1b,c).Element mappings revealed that Pd element was mainly located in the center, while Pt formed the mesoporous shell (Figure 1d).The crystal structure of mPdPt was evaluated by wide-angle X-ray diffraction (XRD) and the XRD pattern showed characteristic diffraction peaks of (111), ( 200), (220), and (311), indicating face-centered cubic (fcc) crystal structure (Figure S4, Supporting Information). [19]The hydrodynamic diameter of mPdPt NPs determined by dynamic light scattering (DLS) was 121 nm, and the zeta potential was À29.3 mV (Figure 1e,f ).
Next, CQ was loaded onto mPdPt NPs (CQ@mPdPt) by suspending mPdPt NPs into different concentrations of CQ and stirred for 12 h.CQ has multiple UV-vis spectra absorption peaks between 200 and 350 nm, and the mixture of free CQ and mPdPt had similar absorption to free CQ (Figure S3, Supporting Information). [20]However, compared to the mPdPt suspension and the mixture of free CQ and mPdPt suspension, CQ-loaded mPdPt showed a broad absorbance between 200 and 350 nm, indicating successful loading of CQ.The Fourier transform infrared (FT-IR) spectra of mPdPt, CQ, and CQ@mPdPt also indicated the loading of CQ as both CQ and CQ@mPdPt had specific bands centered at 1211, 1550, 1612, and 3421 cm À1 , corresponding to the typical bands of C─N, C═C, C═N and N─H stretching vibration of CQ, respectively (Figure S5a, Supporting Information). [21]CQ loading may be through the coordination of metal species with H1 0 or NH atom of the secondary amine in CQ, and physical trapping the molecule into the mesoporous channels. [22]Next, the loading capacity of CQ was optimized.When the mass ratio of mPdPt and CQ was 1:5 (m PdPt ∶m CQ = 1:5), the loading capacity was 16.62%.DLS analysis showed that after CQ loading, the hydrodynamic size of mPdPt increased from 121 to 380.3 nm (CQ@mPdPt), and zeta potential increased from À29.3 to À19.3 mV.To enhance its hydrophilicity and Figure 1.Characterizations of mPdPt, CQ@mPdPt, and TF-CQ@mPdPt.a) SEM images of mPdPt (scale bar: 100 nm).b,c) TEM images of mPdPt (scale bar: 100 nm).d) Element mapping of mPdPt.Green and red represent Pd and Pt, respectively (Scale bar: 100 nm).e) Hydrodynamic size distributions of mPdPt, CQ@mPdPt, and TF-CQ@mPdPt determined by DLS.f ) Zeta potentials of mPdPt, CQ@mPdPt, and TF-CQ@mPdPt.g) SEM images of TF-CQ@mPdPt (scale bar: 200 nm).h) Zeta potential change of TF-CQ@mPdPt over time in different pH conditions.i) Colloidal stability of TF-CQ@mPdPt in water and 10% serum-rich media, respectively.biocompatibility, we coated CQ@mPdPt with a pH-responsive TA-Fe (III) MOF (TF-CQ@mPdPt).TA and Fe (III) could form a coating structure through a self-assembling process by coordination bond between Fe (III) and the pyrogallol group of TA. [23] Compared with mPdPt, the FT-IR spectra of TA-Fe (III) and TF-CQ@mPdPt showed specific centered bands at 1037, 1205, 1533, 1711, and 3360 cm À1 , corresponding to the typical bands of C─C, C─O, C═C, C═O, and O─H stretching vibration of TA, respectively, which indicated the coating of TA-Fe (III) (Figure S5b, Supporting Information). [24]In line with FT-IR spectrum analyses, scanning electron microscope (SEM) images of TF-CQ@mPdPt revealed that there was an obvious coating structure on CQ@mPdPt after TA-Fe (III) coating (Figure 1g).Additionally, the element mappings indicated the existence of Fe and N elements (Figure S6, Supporting Information).Then, mPdPt and TF-CQ@mPdPt were subjected to X-ray photoelectron spectroscopy (XPS) analysis.As shown in Figure S7, Supporting Information, the mPdPt and TF-CQ@mPdPt spectra showed peaks at about 71, 74.9, and 335.5 eV, demonstrating the existence of Pt and Pd. [19]For TF-CQ@mPdPt, there was specific peak located at about 714 eV, indicating the presence of Fe.These results further proved that TA-Fe (III) were successfully coated on CQ@mPdPt.DLS analysis showed that the average size of TF-CQ@mPdPt was 204.6 nm, and the zeta potential was À39.5 mV (Figure 1e,f ).Additionally, under low pH conditions, hydroxyl groups of TA tend to be protonated, thus resulting in a competition between proton hydrogen and Fe (III), leading to the disintegration of the coating structure. [25]To verify the pH-responsive degradation of TA-Fe (III) coating, TF-CQ@mPdPt was dispersed into PBS with different pH values and incubated for 24 h, and then its zeta potentials were determined.Under the condition of pH 5.4, the zeta potential of TF-CQ@mPdPt increased from À41.9 to À21.3 mV after 24 h incubation, while no noticeable change of zeta potential was observed under pH 7.2 (Figure 1h).These observations indicated the pH-responsive disintegration of TA-Fe (III) film.To evaluate the CQ release profile, TF-CQ@mPdPt was dispersed in pH 5.4, 6.4, and 7.2 buffers (1 mg mL À1 ) and then vacillated at the ambient temperature or 45 °C.After incubation for 24 h, the cumulative release of CQ ranged from around 20% to 70% in different conditions (Figure S8, Supporting Information).In consistent with the pH-responsive disintegration of TA-Fe (III), the release of CQ exhibited pH-responsiveness with a gradual release under pH 7.2 and a fast release under the acidic conditions, and the highest release was observed in pH 5.4 buffer at 45 °C.The release of CQ was higher at 45 °C than at 37 °C, revealing that photothermal performance also affected TA-Fe (III) degradation and contributed to CQ release, probably because the increased temperature promoted the thermal motion of molecules.To assess the colloidal stability of TF-CQ@mPdPt in the biological conditions, we recorded the hydrodynamic size change of the particle size in Dulbecco's Modified Eagle Medium (DMEM) containing 10% fetal bovine serum (FBS) at different time periods.Compared to that in water, a sudden increase in size was detected in DMEM (10% FBS), which may arise from the absorption of proteins on the particles (Figure 1i).However, the size did not change over time during 24 h of incubation, indicating good stability of TF-CQ@mPdPt in physiological conditions.

Photothermal and Nanoenzyme Properties of TF-CQ@mPdPt
Next, the photothermal effect and POD-mimic activity of mPdPt and TF-CQ@mPdPt were studied.To evaluate the photothermal performance, TF-CQ@mPdPt was suspended in water with various concentrations (25, 50, 75, 100, 200 μg mL À1 ) and irradiated with a NIR laser (808 nm, 1 W cm À2 ) for 10 min.The temperature of TF-CQ@mPdPt suspension elevated with the increase of its concentration and illumination time, while the temperature of pure water hardly changed (Figure 2a and S9, Supporting Information).As previously reported, the formation of TA-Fe (III) complexes exhibited an intrinsic NIR absorption property which could transfer laser to heat and applicants for photothermal performance. [24]Subsequently, the photothermal performance of TA-Fe (III) and mPdPt was further evaluated according to the content ratio of Fe or Pt in TF-CQ@mPdPt.The temperature rise of TA-Fe (III) suspension was below 6 °C, and the photothermal performance of mPdPt showed an insignificant difference to TF-CQ@mPdPt, which might be due to the much lower content of TA-Fe (III) coating (Pt:Fe % 250:1), so the TA-Fe (III) hardly contribute to the photothermal performance at the test condition (Figure S10 and S11, Supporting Information).Additionally, the photothermal stability of TF-CQ@mPdPt was also evaluated by repeated laser "on-off" exposure (Figure 2b).The temperature of TF-CQ@mPdPt suspension could still recover to its initial maximum level (T % 50 °C) after irradiation for four cycles, indicating that TF-CQ@mPdPt was photothermally stable.To further assess the photothermal conversion effect of TF-CQ@mPdPt, the linear time data versus -lnθ was obtained from the cooling period in Figure 2c, and the photothermal conversion efficiency of TF-CQ@mPdPt was calculated to be 24.7% (Figure S12, Supporting Information). [26]In general, TF-CQ@mPdPt shows good photothermal conversion capacity, similar to some previously reported photothermal agents such as gold nanorods (21%), [27] copper selenide nanocrystals (22%), [28] and Cu 9 S 5 NPs (25.7%). [29]ubsequently, we assessed the POD-mimic activity of mPdPt and TF-CQ@mPdPt.It is known that nanomaterials with POD-mimic activity could catalyze H 2 O 2 to generate •OH and then oxidize 3,3 0 ,5,5 0 -tetramethylbenidine (TMB) to oxTMB (Figure S13, Supporting Information), [30] thus the POD-mimic activity of mPdPt and TF-CQ@mPdPt was evaluated by detecting the UV-vis absorption spectroscopy of oxTMB.TF-CQ@mPdPt and mPdPt were incubated with TMB reagent (containing H 2 O 2 ), and the absorbance of oxTMB at 652 nm was recorded every 10 min.It was found that absorbance of oxTMB demonstrated time-and concentration-dependent manners, indicating the intrinsic POD-mimic activity of mPdPt and TF-CQ@mPdPt (Figure 2d and S14, Supporting Information).As previously reported, TA could reduce the liberated Fe (III) to Fe (II) to produce reactive oxygen species (ROS) through the Fenton reaction. [31,32]However, no obvious absorbance at 652 nm was observed when TA-Fe (III) was incubated with the TMB assay solution (Figure 2d), which was probably due to the much lower content of Fe (III) in TF-CQ@mPdPt, thus the TA-Fe (III) could hardly trigger ROS generation.In addition, the POD-mimic activity of TF-CQ@mPdPt was slightly higher than mPdPt with elongation of incubation time, which may arise from the better colloidal stability of TF-CQ@mPdPt contributed by the TA-Fe (III) coating, as mPdPt may aggregate and settle, thus affecting its enzyme activity over time.Next, electron paramagnetic resonance (EPR) spectroscopy, using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the spin-trapping agent, was conducted to verify the •OH generation.Under mild acid conditions (pH = 5.4), the four-peak spectrum with an intensity ratio of 1:2:2:1, caused by DMPO-OH, was found when TF-CQ@mPdPt was incubated with H 2 O 2 (100 μM), indicating that TF-CQ@mPdPt could trigger the decomposition of H 2 O 2 into •OH (Figure S15, Supporting Information).Meanwhile, methylene blue (MB) can be oxidized into colorless degradation products via intermediate formation by active oxygen species •OH. [31]To explore whether environment pH could affect the POD-mimic performance of TF-CQ@mPdPt, we assessed the degradation of MB after adding H 2 O 2 (10 mM) into TF-CQ@mPdPt suspension in the presence of MB with different pH values (pH = 4, 5.6, 6.4, and 7.2).The absorbance of MB decreased along with the decrease of pH value (Figure 2e), indicating that mild acid environment favors the POD-mimic activity of TF-CQ@mPdPt.Furthermore, to evaluate the enzymatic thermal stability of TF-CQ@mPdPt, we compared the enzymatic activity of TF-CQ@mPdPt with horseradish peroxidase (HRP) under different temperature conditions.Both HRP and TF-CQ@mPdPt had high catalytic activity when the temperature was below 42 °C.However, as the temperature elevated above 50 °C, the relative activity of HRP dramatically decreased from 84.6% to 12.8%, while the enzymatic performance of TF-CQ@mPdPt hardly changed, and its relative activity was 67.6% at 70 °C (Figure 2f ), indicating it had a good catalytic thermal stability in high temperatures and was promising for MPTT.To systematically evaluate the catalytic performance of TF-CQ@mPdPt, we performed a steady-state kinetic assay. [33]First, H 2 O 2 was added into TF-CQ@mPdPt suspensions (37.5 ng mL À1 ) with different concentrations in the presence of TMB (0.5 mM), and time-and H 2 O 2 concentrationdependent absorbance of oxTMB was observed (Figure S16, Supporting Information).Then by analyzing the data according to the Beer-Lambert law to get the Michaelis-Menten saturation curve (Figure S17a, Supporting Information) and Lineweaver-Burk plotting (Figure S17b, Supporting Information), [34] K M and V max were determined to be 10.54 Â 10 À3 M and 1.992 Â 10 À7 M s À1 , indicating the good enzyme activity of TF-CQ@mPdPt.Collectively, these results indicated that TF-CQ@mPdPt exhibited good photothermal performance and POD-mimic activity, which could be used as photothermal agent and catalyst to enhance MPTT.Photothermal performance and POD-mimic activity of mPdPt and TF-CQ@mPdPt.a) Concentration-dependent photothermal effect of TF-CQ@mPdPt irradiated by an 808 nm laser at 1.0 W cm À2 .b) Photothermal stability of TF-CQ@mPdPt (Pt: 75 μg mL À1 ) after laser "on-off " exposure (808 nm, 1.0 W cm À2 ) for four cycles.c) Photothermal profile of TF-CQ@mPdPt (75 μg mL À1 ) after laser exposure (808 nm, 1.0 W cm À2 ).d) Time-dependent absorbance of oxTMB upon the addition of mPdPt (Pt: 75 μg mL À1 ), TF-CQ@mPdPt (Pt: 75 μg mL À1 ; Fe: 300 ng mL À1 ), and TA-Fe (III) (Fe: 300 ng mL À1 ).e) POD-mimic activity of TF-CQ@mPdPt (Pt: 75 μg mL À1 ) in buffers with different pH values.f ) POD-mimic activity of TF-CQ@mPdPt (Pt: 75 μg mL À1 ) and HRP at different temperature.

Intracellular Uptake and Cytotoxicity of TF-CQ@mPdPt In Vitro
To determine the cellular uptake of TF-CQ@mPdPt, 4T1 cells were incubated with TF-CQ@mPdPt (50 and 100 μg mL À1 ) for 4 or 8 h, and then the cells were assessed by transmission electron microscopy (TEM) and the intracellular Pt content was measured by atomic absorption spectrometer (AAS).After incubation with 50 μg mL À1 of TF-CQ@mPdPt for 8 h, TF-CQ@mPdPt was internalized by the cells and located in cell lysosomes (Figure 3a,b).Moreover, the intracellular Pt content increased with increase of nanoparticle concentration and extension of incubation time (Figure 3c), indicating that TF-CQ@mPdPt could be effectively uptaken by the cells.Next, the cytotoxicity of free CQ was assessed.To this end, 4T1 cells were incubated with different concentrations of CQ for 24 h.We found that there was no apparent cytotoxicity of CQ under the conditions examined.Then, we further maintained the CQ-treated cells at 42-45 °C for 20 min to imitate MPTT and found that cell viability decreased in a CQ concentration-dependent manner.However, in the absence of CQ, there was no apparent difference between these two groups, which demonstrated that cell resistance to MPTT and the inhibition of autophagy by CQ could enhance the sensitivity of 4T1 cells to MPTT (Figure 3d).Subsequently, we evaluated the cell viability after the cells were treated with TF@mPdPt, TF-CQ@mPdPt, TF@mPdPt þ laser, and TF-CQ@mPdPt þ laser.4T1 cells were incubated with different concentrations of NPs for 24 h, and laser illumination was performed with an 808 nm laser to maintain the temperature of cell culture media at 42-45 °C for 20 min.The cell viability was assessed by the CCK-8 assay.Both TF@mPdPt and TF-CQ@mPdPt treatments showed a concentration-dependent cytotoxic effect, with TF-CQ@mPdPt being more effective (Figure 3e), possibly because mPdPt with POD-mimic activity produced excessive cytotoxic •OH, and CQ-mediated autophagy inhibition blocked the self-repair of cells.TF@mPdPt þ laser combination treatment was more effective on induction of cell death than TF@mPdPt treatment alone, indicating MPTT could effectively kill cells when coordinating with POD-mimic activity of TF@mPdPt.Notably, TF-CQ@mPdPt þ laser treatment exhibited the most potent cytotoxic effect, suggesting the POD-mimic activity of mPdPt and CQ loaded could significantly enhance MPTT.Therefore, TF-CQ@mPdPt could be used as a photothermal agent to achieve MPTT under laser irradiation, and POD-mimic activity of mPdPt and autophagy inhibition mediated by CQ could significantly enhance the efficacy of MPTT.

Inhibition of Heat Shock Response and Autophagy by TF-CQ@mPdPt In Vitro
ROS could cleave HSP and affect heat shock response by reducing ATP production through damaging mitochondria. [35]To explore whether TF-CQ@mPdPt with POD-mimic activity could induce ROS generation in cells, 4T1 cells were coincubated with TF@mPdPt or TF-CQ@mPdPt (50 μg mL À1 in Pt) for 12 h, and then irradiated with a laser (808 nm, 1.0 W cm À2 ) to maintain the temperature between 42 and 45 °C for 20 min.After the treatments, the cells were further cultured for 6 h and then stained with 2 0 ,7 0 -dichlorodihydrofluorescein diacetate (DCFH-DA), a fluorescent probe which could react with ROS to produce 2 0 ,7 0 -dichlorodihydrofluorescein (DCF), to evaluate intracellular ROS. [36]Confocal laser scanning microscopy (CLSM) observations indicated that compared to the control and MPTT (42-45 °C) groups, cells treated by TF@mPdPt or TF-CQ@mPdPt displayed strong green fluorescence, which indicated that both TF@mPdPt and TF-CQ@mPdPt induced ROS generation in 4T1 cells (Figure 4a).Moreover, TF-CQ@mPdPt treatment showed stronger green fluorescence signals than that by TF@mPdPt, and the distinctions between the laser irradiation group and nonirradiation group were inconspicuous.These results indicated that CQ could magnify the ROS generation and MPTT heating did not affect the POD-mimic activity of mPdPt.It is known that excessive ROS could cause direct oxidative damage to mitochondrial by disrupting electron transfer chain (ETC) function and increasing mitochondrial membrane permeability. [37]Subsequently, we explored whether the NPs could affect mitochondrial function through the oxidative stress induced by ROS.4T1 cells were coincubated with TF@mPdPt or TF-CQ@mPdPt (Pt: 50 μg mL À1 ) in the presence or absence of laser irradiation (808 nm, 1.0 W cm À2 ), and then stained with JC-1 to evaluate the state of mitochondrial.When the mitochondrial membrane potential is high, JC-1 aggregates in the matrix of the mitochondria form polymers (J-aggregates), which produce red fluorescence, while when the mitochondrial membrane potential is low, JC-1 could not gather together and produce green fluorescence as monomers. [38]Therefore, switch of JC-1 from red to green fluorescence can be used as an indicator for detection of mitochondrial damage.Clear red fluorescence from normal mitochondrial could be observed in control and 42-45 °C groups.After treatments by TF@mPdPt, TF-CQ@mPdPt, TF@mPdPt þ laser, and TF-CQ@mPdPt þ laser, the red fluorescence from the treated cells was scarce, but the green fluorescence of depolarized mitochondrial was significantly enhanced, which indicated both TF@mPdPt and TF-CQ@mPdPt could damage mitochondrial (Figure 4b).As energy support for heat shock response, ATP was primarily generated in mitochondrial, and the dysfunctional mitochondrial would reduce intracellular ATP production, thereby impacting heat shock response during MPTT. [39]We further investigated the intracellular ATP after the cells were incubated with TF@mPdPt or TF-CQ@mPdPt in the presence or absence of laser irradiation.Compared to cells in control and 42-45 °C groups, cells treated with TF@mPdPt, TF-CQ@mPdPt, TF@mPdPt þ laser, and TF-CQ@mPdPt þ laser had lower intracellular ATP content (Figure 4c), which decreased 14.1%, 8.29%, 15.1%, and 9.36%, respectively, after the treatments.These results indicated that mitochondrial damage led to decrease of ATP production and CQ could enhance the damage to mitochondrial.
Next, we further assessed the expression of HSP70 after MPTT to test whether mPdPt could inhibit the heat shock response.4T1 cells were incubated with TF@mPdPt (50 μg mL À1 in Pt) for 12 h, and then irradiated by laser (808 nm, 1 W cm À2 ) to heat the cell culture media to 42-45 °C and maintain for 20 min.Compared to the cells maintained under the normal condition, cells cultured under 42-45 °C condition (20 min, positive control) displayed strong green fluorescence, indicating the upregulation of HSP70.TF@mPdPt treatment hardly induced upregulation of HSP70; however, TF@mPdPt could markedly suppress HSP70 expression after MPTT (Figure 4d).Subsequently, a western blotting assay was further conducted to reveal intracellular HSP70 expression levels after different treatments.4T1 cells were incubated with TF@mPdPt or TF-CQ@mPdPt, and then irradiated by laser (808 nm, 1 W cm À2 ) to maintain the cell culture media to 42-45 °C for 20 min.The untreated cells cultured under 42-45 °C for 20 min were used as positive control.Compared with that in the positive control (42-45 °C) group, HSP70 expression in cells treated with TF@mPdPt þ laser and TF-CQ@mPdPt þ laser was significantly decreased, with TF-CQ@mPdPt being more effective (Figure 4e,f ), which was consistent with the results of immunofluorescence staining of HSP70.These results indicated TF@mPdPt and TF-CQ@mPdPt could suppress the HSP70 expression after MPTT.Collectively, these results suggested that MPTT could trigger heat shock response with upregulated HSPs, while combined treatment with TF-CQ@mPdPt could effectively inhibit the upregulation of HSPs through ROS production and energy depletion, thus enhancing the effect of MPTT.
CQ is a common inhibitor of autophagy. [40]By disturbing lysosome function, it could restrict fusion of initial autophagic vacuoles (AVi) with lysosome to form degradative autophagic vacuoles (AVd), which blocks cell autophagy. [41]To verify whether autophagy could be induced by MPTT and then inhibited by CQ, 4T1 cells were incubated with TF@mPdPt or TF-CQ@mPdPt (Pt: 50 μg mL À1 ) for 12 h, and laser irradiation (808 nm, 1 W cm À2 ) was conducted to maintain the temperature of cell culture media at 42-45 °C for 20 min.Then, bio-TEM examinations of 4T1 cells were performed to investigate intracellular autophagic flux after different treatments (Figure 5a).Cells in the control group were featured with few AVi and AVd, indicating that the autophagy was canonical.However, increased AVi and AVd were observed in TF@mPdPt, TF-CQ@mPdPt, TF@mPdPt þ laser, and TF-CQ@mPdPt þ laser groups, indicating the autophagy was induced after treatments.Moreover, compared to TF@mPdPt and TF@mPdPt þ laser groups, more accumulated AVi were observed in TF-CQ@mPdPt and TF-CQ@mPdPt þ laser groups, which suggested that the CQ released from TF-CQ@mPdPt caused lysosomal dysfunction in 4T1 cells and blocked fusion of AVi with lysosome.Moreover, microtubule-associated protein 1A/1B light chain 3 (LC3) is vital for autophagy.When autophagy was induced, LC3 I would convert to LC3 II. [42]Therefore, detecting LC3 I and LC3 II could be a reliable way to monitor autophagy occurrence.Also, the specific polyubiquitin-binding protein p62 would be recruited to AVi and then degraded in AVd during autophagy, which can be used to determine the inhibition of autophagy by its accumulation. [43]ubsequently, western blotting assay of 4T1 cells were performed to evaluate the cellular expression levels of LC3 I, LC3 II, and p62 after different treatments (Figure S18, Supporting Information).Consistent with bio-TEM observations, cells treated with TF@mPdPt þ laser and TF-CQ@mPdPt þ laser showed increased levels of LC3 II, indicating autophagy occurred after the treatments (Figure 5b).Notably, compared with the cells in TF@mPdPt þ laser, those in TF-CQ@mPdPt þ laser had accumulated more p62 (Figure 5c), which revealed that the CQ released from TF-CQ@mPdPt inhibited the fusion of AVi with lysosomes, resulting to the autophagy inhibition of 4T1 cells.In general, compared to TF@mPdPt group, TF-CQ@mPdPt treatment induced more ROS generation, mitochondria damage, energy loss, and Avi accumulation, indicating the CQ loading imparts mPdPt with the suppressive function of autophagy.Collectively, these results demonstrated that MPTT induced heat shock response and autophagy in tumor cells.TF-CQ@mPdPt could inhibit the expression of HSP70 by enhancing intracellular •OH production and limiting ATP energy supply, while CQ released from TF-CQ@mPdPt blocked the fusion of AVi and lysosome to inhibit autophagy.Simultaneous inhibition of heat shock response and autophagy would sensitize tumor cells to MPTT.

Tumor Targeting and Biodistribution of TF-CQ@mPdPt
The tumor-targeting efficiency of TF-CQ@mPdPt was evaluated using a multispectral optoacoustic tomography (MSOT) system due to its good photothermal effect in the NIR region.When the tumor volume reached %100 mm 3 , 4T1 tumor-bearing mice were intravenously injected with TF-CQ@mPdPt (Pt: 10 mg kg À1 b.w.), and the photoacoustic imaging (PAI) were performed at different time points after injection.The PAI signals in the tumor area were observed at 1 h postinjection, and the signal intensity increased over time and reached the maximum at 12 h after injection (Figure 6a and S19, Supporting Information), indicating that Figure 6.Tumor targeting and biodistribution of TF-CQ@mPdPt.a) Representative PAI images of tumors after the tumor-bearing mice were intravenously injected with TF-CQ@mPdPt (Pt: 10 mg kg À1 ).b) Distribution of TF-CQ@mPdPt (in %ID/g of Pt) in major organs after 12 h intravenous injection of TF-CQ@mPdPt.c) Real-time thermal IR images of tumor-bearing mice after intravenous injection of TF-CQ@mPdPt under 808 nm laser irradiation (0.755 W cm À2 , 20 min) (error bars represent mean AE SD (n ≥ 3)).
TF-CQ@mPdPt gradually accumulated in the tumor and culminated at 12 h after injection, which could be the optimal time point for MPTT.Subsequently, biodistribution of TF-CQ@mPdPt (in %ID/g) was determined 12 h after intravenous injection.TF-CQ@mPdPt was primarily sequestered by liver and spleen, and tumor had a moderate accumulation (Figure 6b).To determine the optimal laser power density for tumor MPTT, tumorbearing mice were intravenously injected with TF-CQ@mPdPt (Pt: 10 mg kg À1 b.w.), and 12 h after injection, tumors were irradiated with an 808 nm laser at different laser power (Figure S20a, supporting information).When illuminated at 0.755 W cm À2 for 15 min, the tumor temperature increased to around 45 °C and kept constant even with further irradiation.Therefore, the laser power for tumor MPTT was determined to be 0.755 W cm À2 .Then, the photothermal images of tumors were recorded to monitor temperature elevation in real time (Figure 6c, S20b, Supporting Information).The tumor temperature of the TF-CQ@mPdPt-treated mice increased to 42 °C after 5 min irradiation and could remain at 42-45 °C under irradiation.However, for the control mice, temperature elevation in tumor region was marginal under the same laser irradiation condition and the temperature was lower than 40 °C after 20 min of irradiation.These results revealed that TF-CQ@mPdPt could target tumors and effectively transfer light to local heat, which made it promising to act as a photothermal agent for tumor MPTT.
2.6.Enhanced Antitumor Effect of TF-CQ@mPdPt-Mediated MPTT In Vivo Next, we investigated whether TF-CQ@mPdPt could enhance MPTT on 4T1 tumors in vivo.When tumors grew to around 100 mm 3 , 4T1 tumor-bearing mice were randomly divided into six groups: control, laser (0.755 W cm À2 , 20 min), TF@mPdPt (Pt: 10 mg kg À1 ), TF-CQ@mPdPt, TF@mPdPt þ laser, and TF-CQ@mPdPt þ laser, respectively (Figure 7a).One day after the treatments, HSP70 and p62 staining of tumor tissues was carried out to explore heat shock response suppression and autophagy inhibition.Compared with the control group, the enhancement of HSP70 fluorescence in TF@mPdPt þ laser group was stronger than in TF-CQ@mPdPt þ laser group, indicating the upregulation of HSP70 after MPTT and the HSP70 expression suppression by TF-CQ@mPdPt, while tumor tissues of mice treated with laser alone did not exhibit evident HSP70 fluorescence enhancement, possibly because the low local temperature (%39 °C) could hardly trigger strong heat shock response (Figure S21, Supporting Information).In addition, compared with the control group, tumor tissues treated with CQ, TF-CQ@mPdPt, and TF-CQ@mPdPt þ laser displayed stronger green p62 fluorescence, indicating the accumulation of p62 and the inhibition of autophagy (Figure S22, Supporting Information).Subsequently, the tumor volumes of the mice in each group were measured every 2 days for 14 days.Tumors in control group grew rapidly, while a slight tumor inhibition was observed in laser and TF@mPdPt groups, indicating that a single treatment was not effective enough for tumor inhibition (Figure 7b,c).TF@mPdPt þ laser and TF-CQ@mPdPt treatments displayed a moderate inhibition effect on tumor growth, which was probably due to the synergistic anticancer effect of POD-mimic activity and MPTT for group TF@mPdPt þ laser, and POD-mimic activity and autophagy inhibition of CQ for group TF-CQ@mPdPt.Notably, TF-CQ@mPdPt þ laser treatment showed a higher tumor inhibition effect than any other treatment, and almost completely suppressed the tumor growth within 14 days, indicating the enhanced therapeutic effect of MPTT by POD-mimic activity and CQ.Furthermore, the hematoxylin and eosin (H&E) staining of tumor tissues was performed 2 days after different treatments to verify the therapeutic effect (Figure 7d).Compared with control group, varying degrees of cell damage was observed in all treatment groups, and the most severe damage was observed in the tumor tissues from TF-CQ@mPdPt þ laser group, consistent with its most effective tumor inhibition effect.Simultaneously, a terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay was conducted to detect cell apoptosis (Figure 7d).The green fluorescence from apoptotic cells was evident in all treatment groups and strongest in TF-CQ@mPdPt þ laser group, revealing that TF-CQ@mPdPt þ laser was the most effective strategy to inhibit tumor growth.
Finally, to assess the biosafety of our treatment strategy, mice weights in all treatment groups were recorded every 2 days within the treatment period (Figure 7e).No notable weight fluctuations were observed in all groups, which indicated that the adverse effect of those treatments was negligible.We further evaluated the biosafety of the treatments by H&E staining of the major organs (including hearts, livers, spleens, lungs, and kidneys) collected after treatments.Compared with the control group, unnoticeable damage in major organs was observed (Figure S23, Supporting Information).These results demonstrated that our treatment strategy was biocompatible and safe, which was vital for its application in tumor MPTT.

Conclusion
In this study, we successfully developed a tannic acid-iron ion metal organic framework-coated, CQ-loaded mPdPt nanosystem (TF-CQ@mPdPt) to enhance MPTT by simultaneous suppression of heat shock response and autophagy.TF-CQ@mPdPt exhibited good photothermal conversion performance and POD-mimic activity.Upon tumor accumulation, the mPdPt-mediated POD-mimic catalytic reaction would suppress the upregulation of HSPs during MPTT by decomposition of the endogenous H 2 O 2 to generate excessive ROS, leading to improved sensitivity of tumors to heat stress.CQ released from TF-CQ@mPdPt inhibited autophagy and blocked the self-repair process of tumor cells, thus further enhancing the effect of MPTT.MPTT in combination with TF-CQ@mPdPt demonstrated effective antitumor effect both in vitro and in vivo, highlighting that simultaneous inhibition of heat shock response and autophagy was a promising strategy to enhance the therapeutic effect of MPTT.