A cyclic nano-reactor achieving enhanced photodynamic tumor therapy by reversing multiple resistances

Background Photodynamic therapy (PDT) is a clinically implemented modality to combat malignant tumor, while its efficacy is largely limited by several resistance factors from tumor microenvironment (TME), such as hypoxia, anti-oxidant systems, and ATP-dependent tumor adaptive resistances. The aim of this work is to construct a multifunctional nanoplatform to remodel multiple resistant TME for enhanced PDT. Results Here, a targeting nano-reactor was facilely constructed to reverse the multiple resistances of PDT by incorporating glucose oxidase (GOx) and chlorin e6 (Ce6) into poly (D, L-lactic-co-glycolic acid) (PLGA)/ metalorganic framework (MOF) coreshell nanoassembly, with surface deposition of hyaluronic acid (HA) stabilized MnO2. The nano-reactor could selectively target tumor cells by virtue of surface HA modification, and once internalization, a few reactions were initiated to modulate TME. Glucose was consumed by GOx to inhibit ATP generation, and the produced H2O2 was catalyzed by MnO2 to generate O2 for tumor hypoxia alleviation and photodynamic sensitization, and glutathione (GSH) was also effectively depleted by MnO2 to suppress the tumor antioxidant defense. Consequently, the nano-reactor achieved robust PDT with amplified tumor therapy via intravenous injection. Conclusions This nano-reactor offers a multifunctional nanoplatform to sensitize TME-limited tumor treatment means via reversing multiple resistances. Grpahic abstract Supplementary Information The online version contains supplementary material available at 10.1186/s12951-021-00893-6.

techniques, tumor resistances are the main mechanism to weaken the efficacy of PDT [11].
Tumor cells can resist PDT by different pathways due to the complexity of tumor microenvironment (TME), in which the most well-known feature is hypoxia. Hypoxic cells are ~ threefold more resistant to ROS damage than aerobic cells [12], and more importantly, hypoxic environment could directly decrease the PDT efficacy by blocking the oxygen supply [13,14]. What's more, the PDT process would further aggravate tumor hypoxia through oxygen consumption and vascular damage, which activates hypoxia inducible factor-1α (HIF-1α) survival pathway, thus causing PDT resistance [15,16]. At the same time, cancer cells are equipped with antioxidant defense systems, in which the most abundant one is glutathione (GSH), to scavenge the ROS and thus counteract ROS-mediated injury [17][18][19]. In addition, tumor cells could generate adaptive resistance toward PDT through upregulation of drug efflux proteins, heat shock proteins (HSPs), and DNA repair proteins [20][21][22]. It has been reported that numerous drug efflux proteins, such as P-glycoprotein (P-gp), and ATP-binding cassette super-family G member 2 (ABCG2), have been implicated in PDT resistance via pumping out PSs before their action [23].
To reverse PDT resistance, various nano-systems capable of modulating TME have been developed. For example, replenishment of oxygen is a commonly employed strategy to alleviate hypoxia, which can be achieved by either oxygen delivery (using hemoglobin or perfluorocarbons) or endogenous oxygen generation (through catalytic converting tumor abundant H 2 O 2 into O 2 ) [24][25][26]. Oxygen generation can be further boosted by cascade equipping the nano-system with glucose oxidase (GOx) to supplement H 2 O 2 substrate [27]. To augment PDT efficacy, several nano-vehicles were also designed to suppress the tumor antioxidant defense and restore the ROS damage effect of PDT via antioxidants depletion [28,29]. Moreover, inhibition of ATP was reported to sensitize PDT by inhibiting drug efflux and aggravating PDTinduced DNA damage [30,31]. Ideally, simultaneous remodeling various TME to alleviate multiple biological resistances is preferred for optimized PDT, considering the cunning nature of cancer. Unfortunately, the development of nanoparticles with multifunctionalities often requires sophisticated materials design and complicated preparation procedure, thus imposing the difficult of cost-effective, reproducible, and scalable production.
To tackle this, we designed and facilely prepared a core-shell nano-reactor for enhanced PDT via simultaneous oxygenation, antioxidant suppression, and ATP depletion (Scheme 1). Tannic acid (TA), ferric iron (Fe 3+ ) and poly (D, L-lactic-co-glycolic acid) (PLGA) were assembled into core-shell structure via hydrophobic, π-π staking, and electrostatic interactions [32][33][34], in which chlorin e6 (Ce6, a widely used PS) and GOx were Scheme 1 Schematic illustration the preparation of core-sell nano-reactor, and its targeted delivery in vivo for enhanced PDT against tumor by reversing multiple resistances co-loaded. MnO 2 was deposited on particle surface and stabilized by hyaluronic acid (HA) through layer-by-layer electrostatic adsorption. In our system, the MnO 2 played dual roles of self-oxygen supply and GSH depletion via its catalase-mimic activity and oxidbillity, respectively. The GOx boosted the oxygenation by in-situ generation of H 2 O 2 , and its capability of glucose consumption blocked the energy supply to decrease ATP generation. All these functions collectively remodeled the TME from different aspect to sensitize PDT efficacy, which has been demonstrated by solution, in vitro and in vivo experimental results. The nanosystem could passively accumulate into tumor, actively internalize tumor cells via HA-mediated targeting, and impose potent PDT effect via reversing multiple resistances.

Preparation of PTFCG@MH
Ce6, FeCl 3 and PLGA were mixed in acetone at a concentration of 400 μg/mL, 400 μg/mL and 2 mg/mL, respectively. Then, 1 mL of such mixture was added dropwise to 5 mL aqueous solution containing 320 μg/mL TA and 20 μg/mL GOx under constant sonication. The solution was stirred at 30 °C for 3 h to evaporate acetone. The PTFCG was separated by centrifugation (16,000 rpm, 25 min) and washed by deionized water. To prepare PTFCG@MH, 50 μL PAH (20 mg/mL) was mixed with 1 mL PTFCG solution and mild stirred for 30 min. Then, 25 μL potassium permanganate (10 mg/mL) was added and stirred for another 1 h. After centrifugation (16,000 rpm for 25 min) and washing with deionized water, the PTFCG@M was obtained. Finally, 500 μL HA (20 mg/mL) was introduced into 1 mL of PTFCG@M solution and stirred for 1 h. Excess HA was removed by centrifugation (16,000 rpm, 25 min), and PTFCG@MH was collected.

Characterizations of PTFCG@MH
The particle sizes and ζ potential of PTFCG, PTFCG@M and PTFCG@MH were measured by a Malvern Zeta Sizer Nano series (Malvern, UK) at 25 °C. The morphologies and element content were investigated by TEM-EDS (Titan G2 60-300, FEI, USA). For TEM-EDS analysis, the PTFCG or PTFCG@MH solution was dropped onto carbon film-coated copper grid and observed under TEM-EDS instrument. The UV-vis absorption spectra were carried out on UV-visible spectrophotometer (Shimadzu, Japan). The fluorescence spectra were collected using FL-2700 spectrofluorometer (HITACHIH). The Ce6 concentration was determined by the UV-visible spectra spectrophotometer at 640 nm. The loading amount of GOx was estimated using bicinchoninic acid (BCA) protein assay (Beyotime) and the protocol was provided by the supplier. The drug loading (DL) of Ce6 and GOx in the PTFCG@MH was calculated as follows:

Cyclic reaction of PTFCG@MH
The catalytic activity of GOx was determined by measuring the glucose consumption and pH change in presence of glucose. For the glucose consumption, the PTF@MH or PTFCG@MH solution ([GOx] = 4 μg/mL) was mixed with glucose (10 mM) in the presence or absence of H 2 O 2 DL = amount of loaded drug in the PTFCG@MH / weight of PTFCG@MH (µg/mg).

Cellular uptake
The NR-loaded nanoparticles were prepared following the method described above by replacing Ce6 with NR. The cells were seeded in Petri dish (35 mm) at 5 × 10 4 cells/cm 2 and cultured overnight. With or without pretreatment of free HA (10 mg/mL) for 1 h, the cells were incubated with NR-loaded nanoparticles or free NR ([NR] = 2 μg/mL) for another 2 h. Next, the cells were washed with pre-cooled PBS and fixed with 4% paraformaldehyde. After staining by DAPI (1 μg/mL), the cells were imaged under confocal microscope (LSM780 NLO, Zeiss, Germany).

Intracellular O 2 consumption and ROS detection
ROS-ID and DCFDA were used to detect the intracellular O 2 consumption and ROS generation inside cells. The MDA-MB-231 cells were seeded in 24-well culture plate (10 5 cells/well) and cultured overnight. After sealing by liquid paraffin, the cells were incubated with PTFCG@ MH, PTF@MH, PTFCG, or free Ce6 ([Ce6] = 1 μg/mL) for 2 h. Then, laser irradiation (635 nm, 100 mW/cm 2 ) was applied for 2 min. With PBS washing for three times, the mixture of ROS-ID (0.5 μM) and DCFDA (10 μM) was added into each group for 30 min further incubation. After washing with pre-cooled PBS, fluorescence inside cells was observed by fluorescence imaging system (NIKON, Ti-S, Japan).

Intracellular ATP and GSH detection
MDA-MB-231 cells were seeded in 12-well plate at a density of 2 × 10 5 cells per well for 12 h incubation. Then, various formulations with different concentrations were added for 24 h incubation. After that, the cells were collected and lysed using ATP lysis buffer, and ATP concentration was measured by ATP assay kit (Beyotime). Likewise, the cells were collected and lysed using the Triton-X 100 cell lysis buffer, and GSH concentration was detected by the Ellman's reagent (Solarbio). The numbers of cells were standardized by measuring total protein concentration using BCA protein assay (Beyotime).

In vitro cytotoxicity studies
MDA-MB-231 cells were seeded in a 96-well plate (5000 cells/well) for overnight incubation. The cells were treated with PTFCG or PTFCG@MH at different concentrations for 24 h, followed by laser irradiation for 1 min (100 mW/cm 2 , 635 nm). After 24 h, MTT solution (1 mg/mL) was added for another 4 h incubation. The medium was replaced with 100 μL of DMSO. The absorbance values at 570 nm were recorded to calculate the cell viability. Then, the Live/Dead assay was carried out by calcein AM/propidium iodide double staining. The cells were seeded and treated as described above. After laser irradiation (100 mW/cm 2 , 635 nm), the fresh medium containing calcein AM (2 μM) and propidium iodide (8 μM) was added for 20 min incubation. Finally, the cells were imaged by fluorescence imaging system (NIKON, Ti-S, Japan).

In vivo/ex vivo fluorescence imaging
Healthy female Balb/c mice (aged 4-6 weeks) were purchased from the Laboratory Animal Center of Central South University. All experimental procedures were carried out in accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals of China, and approved by the Ethics Committee for Research in Animal Subjects at Xiangya School of Pharmaceutical Sciences of Central South University. MDA-MB-231 tumor-bearing mice were obtained by subcutaneously injecting a cells suspension in PBS (10 6 cells) into the right armpit of mice. When the tumor volume reached ~ 100 mm 3 , the mice were treated with different formulations.
For in vivo/ex vivo imaging, the tumor-bearing mice were intravenously injected with free Ce6 or PTFCG@ MH (100 μL, [Ce6] = 2.5 mg/kg), and the fluorescence images were taken by an optical imaging system (IVIS Lumina, PerkinElmer, USA) at 1 h and 24 h post-injection. The mice were sacrificed at 24 h after injection, and the major organs were collected for ex vivo imaging. The images were analyzed using Living Imaging Software (IVIS Lumina LT, PerkinElmer, USA).

In vivo antitumor study
The tumor-bearing mice were randomly divided into 5 groups and injected intravenously with (1) PBS, (2) Ce6 + Laser, (3) PTFCG + Laser, (4) PTFCG@MH, (5) PTFCG@MH + Laser (100 μL, [Ce6] = 2.5 mg/kg) at day 0 and 4. The irradiation groups were exposed to laser (100 mW/cm 2 , 635 nm) for 5 min at 24 h after injection. The tumor volume was recorded and calculated as follows: V = (length × width 2 )/2. In addition, the body weights were also obtained. At day 14, all the mice were sacrificed, and the tumors were collected and weighed. For histology analysis, the major organs and tumors were extracted and immersed in 4% formaldehyde, embedded in paraffin, sectioned, stained with H&E and observed by an optical microscope (Leica, German).

Immunofluorescence analysis
The expression of HIF-1α protein was investigated by immunofluorescence assay. Nude mice bearing MDA-MB-231 tumors were sacrificed at 48 h after various treatments mentioned above. Tumor tissues were collected for preparing slices of tumor frozen sections. Then, the slices were stained with HIF-1α antibody (mouse polyclonal to HIF-1α, Abcam) at 4 °C overnight, followed by adding the secondary antibodies conjugated with FITC for 1 h incubation. After cell nuclei staining with Hoechst 33,258 for 20 min, the tumor sections were imaged by confocal fluorescence microscope (LSM780 NLO, Zeiss, Germany).

Statistical analysis
Data were presented as mean ± SD. Analysis was performed using Graphpad Prism 5 software. One-way ANOVA analysis of variance was conducted to determine the statistical significance of various groups. P values < 0.05 was regarded as statistically significant.

Nanoparticles preparation and characterization
The GOx and Ce6 were co-loaded into nanocomposites (PTFCG) using a solvent exchange and evaporation method (Scheme 1), in which the ethanol solution (containing Ce6, PLGA and FeCl 3 ) was dropwise added into the aqueous phase (with TA and GOx). During organic evaporation, the TA-Fe metal-organic frameworks (MOFs) were coated on the surface of hydrophobic PLGA nano-core for particle stabilization, and the resulting PTFCG displayed a dynamic diameter of ~ 175 nm (Fig. 1A, Additional file 1: Figure S1) with ζ potential of − 33.4 mV (Fig. 1C). From the TEM image, the PTFCG displayed a roughly spherical morphology with an obvious core-shell structure (inset in Fig. 1A).
To further deposit MnO 2 on particle surface, a PAH solution was added, followed by adding KMnO 4 , allowing for in-situ growth of MnO 2 on nanoparticle surface to yield PTFCG@M. Compared with PTFCG, the ζ potential of PTFCG@M was reversed to positive (+ 17.3 mV) (Fig. 1C), which allowed for the subsequent HA coating via electrostatic attraction (termed PTFCG@MH). The successful HA modification was evidenced by the decrease of particle charge to negative (− 21.7 mV) (Fig. 1C), which is beneficial for enhanced colloidal stability. [35] The resulting PTFCG@MH nanoparticles displayed a dynamic diameter of ~ 205 nm (Fig. 1B), while the size measured by TEM was ~ 160 nm (inset in Fig. 1B, Additional file 1: Figure S2). It is reasonable as DLS measurement would distort the particle size being observed due to the particle surface hydration. Importantly, the HA modification not only significantly increased the colloidal stability (Additional file 1: Figure S3), [36] but also rendered the nanoparticles with tumor targetability (vide infra).
According to the UV-vis absorbance spectra, PTFCG had characteristic Ce6 peaks at 404 nm and 640 nm, indicating the successful Ce6 loading (Fig. 1D). For PTFCG@ MH, a wide absorbance band in the range from 250-400 nm appeared, which was ascribed to the MnO 2 . To confirm the MnO 2 coating, the energy dispersive X-Ray spectroscopy (EDS) was conducted, and a high Mn element content (~ 6%) was observed (Fig. 1E). The drug loading (DL) for Ce6 and GOx was measured to be 108.4 μg and 24.6 μg per mg of PTFCG@MH, respectively. Interestingly, after encapsulation into nanoparticles, the intrinsic fluorescence of Ce6 was significantly quenched, especially for PTFCG@MH (Additional file 1: Figure S4). However, such quenched fluorescence can be largely recovered upon addition of GSH, indicating GSHresponsive release of Ce6. To confirm this, Ce6 release behavior from PTFCG@MH was investigated (Fig. 1F). In absence of GSH, the Ce6 showed ~ 20% accumulative release, while a burst drug release was observed after adding GSH, with ~ 60% Ce6 release within 8 h. Therefore, such nanosystem showed a triggered activation of photodynamic activity in response to intracellular stimulus, thus minimizing the potential phototoxicity of Ce6 during in vivo delivery.

A cyclic reaction nano-reactor
As illustrated in Fig. 2A, our nanosystem was designed with a few cyclic reactions to enhance the Ce6-based PDT. GOx consumes glucose (Glu) to block the energy (ATP) supply, and the concomitantly produced H 2 O 2 is decomposed by the catalase-mimic MnO 2 to generate O 2 , which in turn boosts the 1 O 2 production under laser illumination. To demonstrate this concept, each reaction was tested individually.
The catalytic activity of GOx was first measured by monitoring the glucose consumption (Fig. 2B). After 60 min incubation, PTFCG@MH decreased the glucose level by 34%, while the nanoparticles without GOx loading (termed PTF@MH) did not show any glucose consumption. Therefore, GOx maintained its catalytic activity after being loaded into nanoparticles. Moreover, the glucose depletion of PTFCG@MH was further strengthened after adding H 2 O 2 , which can be ascribed to the decomposition of H 2 O 2 by PTFCG@MH to supply O 2 . Michaelis-Menten constant (K M ) was calculated to quantify the reaction kinetics, and the K M value of PTFCG@MH + H 2 O 2 and PTFCG@MH turned out to be 6.35 and 10.02 mM (Additional file 1: Figure S5), indicating the promoted glucose consumption by H 2 O 2 . The activity of GOx can also be monitored by the pH decrease due to the generation of gluconic acid byproduct. Upon addition of glucose, both PTFCG and PTFCG@MH showed a gradual pH decrease over time (Fig. 2C). Notably, PTFCG@MH exhibited a relative lower pH decrease rate than that of PTFCG. This is likely due to that the surface deposited MnO 2 impeded accessibility of glucose into GOx-loaded nanocore. In addition, the MnO 2 could neutralize H + to slow down pH decrease rate.
Next, the catalase-mimic activity of MnO 2 was tested by monitoring O 2 production in presence of H 2 O 2 (Fig. 2D). As a control, the PTFCG@MH alone did not show any O 2 generation. However, a rapid increase of the dissolved O 2 was observed upon addition of H 2 O 2 . We also studied the effect of pH, and a marginal increase of catalytic rate was seen with the decrease of pH from 7.0 to 5.0. Overall, the catalytic efficiency was relatively higher under acidic tumor microenvironment than physiological conditions.
After confirming the single catalytic reaction of GOx and MnO 2 , we next explored the catalytic circulation by measuring the oxygen variation (Fig. 2E). For PTFCG, the dissolved O 2 was quickly decreased within 100 s upon addition of glucose. For the PTFCG@MH, by contrast, the O 2 consumption rate significantly lessened due to the self-oxygen generation activity of MnO 2 , demonstrating the cyclic oxygen supply. Moreover, the O 2 balance was achieved upon further addition of 100 μM H 2 O 2 (which mimics the tumor microenvironment) [37], and such O 2 generation efficiency is strong enough to support the PDT (Fig. 2E, blue trace).
With cyclic oxygen supply, we next explored the enhanced PDT effect by measuring the single oxygen ( 1 O 2 ) generation using singlet oxygen sensor green (SOSG) as a fluorescent indicator (Fig. 2F). In presence of H 2 O 2 , the production of 1 O 2 was obviously increased for PTFCG@MH, and the addition of glucose could also strengthen the PDT effect. As a control, the PTFCG group showed lower 1 O 2 generation due to the lack of oxygen supply. All these results demonstrated the capability of cyclic nano-reactor for boosting PDT effect of Ce6.

Cellular uptake
We next tested the intracellular performance of the nanosystem by using MDA-MB-231 cancer cells. To track the intracellular delivery, nanoparticles were labeled with a red fluorescent Nile red (NR), and the cell nuclei were stained blue by DAPI for localization (Fig. 3A). From confocal laser scanning microscopy (CLSM) images, a weak red signal was observed when the cells were treated with free NR, while the fluorescence was significantly intensified for nanoparticles. From the merged image, the fluorescence of nanoparticles mainly distributed in the cytoplasm, indicating endocytosis pathway for internalization. To demonstrate tumor targetability of such surface HA modified nanosystem, the cells were pretreated with free HA to saturate the CD44 binding, and in this case the intracellular nanoparticles signal decreased obviously. We also quantified the contribution of HA-mediated internalization by measuring the intensity of each treatment, where the uptake was significantly decreased upon free HA pre-treatment (Fig. 3B). Using normal human umbilical vein endothelial cells (HUVEC) as control, the results further demonstrated the targetability of the nanoparticles towards tumor cells with significantly higher fluorescence (Additional file 1: Figure S6). These results confirmed that our nanosystem was able to selectively recognize tumor cells via CD44 receptor for targeting delivery. To further explore the delivery mechanism, various endocytosis inhibitors were employed, including chlorpromazine, nystatin and colchicine, which blocks clathrin-mediated endocytosis, caveolae-mediated endocytosis and micropinocytosis, respectively. Among then, chlorpromazine has the most significant impact on nanoparticles uptake (Fig. 3C, D), indicating the main contribution of clathrin-mediated endocytosis.

Enhanced anti-tumor efficacy via hypoxia alleviation and ATP/GSH depletion
After internalization, we next explored the intracellular functions of the nanosystem. To visualize the PDT effect, the ROS generation was probed by using 2′ 7'-dichlorofluorescin diacetate (DCFDA) indicator. Interestingly, both PTFCG@MH and PTF@MH could scavenge the intracellular ROS to some extent as compared to the control, due to the catalase activity of MnO 2 . Upon irradiation, each Ce6-based formulation showed enhanced green fluorescence inside cells based on both fluorescent images (Fig. 4A) and quantified intensity (Fig. 4B). Notably, PTFCG@MH group emitted the strongest fluorescence, indicating the best PDT efficacy. This can partially be explained by the cyclic oxygenation of the nano-reactor for self-oxygen supply. To confirm this, we then measured the O 2 balance by using a red fluorescent hypoxia detection kit. Both PTF@MH and PTFCG@ MH could relieve tumor hypoxia with red fluorescence decrease compared to non-treatment control (Fig. 4A, C), also attributable to self-oxygen supply of the MnO 2 . Upon addition of laser, the hypoxia was strongly exacerbated for free Ce6 and PTFCG groups, ascribed to PDTand GOx-based oxygen consumption. However, the PTFCG@MH group showed low hypoxia level even after laser irradiation. Therefore, such nano-reactor is robust enough to maintain oxygen balance with enhanced PDT efficacy.
Besides cyclic oxygen supply, such nano-reactor was also designed to deplete both ATP and GSH. We next tested these properties by measuring the intracellular ATP and GSH levels using ATP assay kit and Ellman's reagent, respectively. Because of catalytic glucose oxidation, free GOx could effectively decrease ATP level in a concentration dependent manner by blocking the energy supply (Fig. 4D). Likewise, the PTFCG@MH could also inhibit ATP generation, while the PTF@MH did not show any activity due to the absence of GOx loading. MnO 2 is a well-known GSH depletor due to its capability to oxidize GSH into GSSG, accompanied by its reduction into Mn 2+ [28,38]. As expected, the PTFCG@MH with MnO 2 doping displayed high efficient GSH depletion, with over 80% GSH decrease at 5 μg/mL nanoparticles (Fig. 4E). Note that ATP is the basic energy source for tumor cells to acquire treatments resistance including PDT therapy [39], and GSH could directly scavenge 1 O 2 to alleviate the PDT efficacy. Therefore, with ATP and GSH dual-depletion activity, such nano-reactor was expected to enhance the PDT-based anti-tumor activity via distinct mechanisms.
Next, the in vitro cytostatic activity of nanosystem was evaluated by MTT assay. Without drugs loading, the PTF and PTF@MH showed satisfactory biocompatibility even at high concentrations (Additional file 1: Figure  S7). The PTFCG, on the other hand, displayed a concentration-dependent tumor ablation activity (Fig. 4F), ascribed to the starvation therapy for ATP depletion [40]. Upon laser irradiation, the antitumor effect was further enhanced. Notably, PTFCG@MH showed significantly better efficacy than PTFCG attributable to MnO 2 doping for self-oxygen circulation, GSH depletion, as well as HA modification for targeting delivery. To confirm the tumor selectivity, MTT assay was also performed on HUVEC  Figure S8). We also explored the cell death pathway by co-staining the cells using calcein-AM (green fluorescence for live cells) and propidium iodide (red fluorescence for dead/late apoptotic cells). After different treatments, green fluorescence was weakened while the red fluorescence became intensified inside cells (Fig. 4G), indicating an apoptotic or necrotic cell death mechanism. In addition, the general anti-tumor activity for each treatment was highly consistent with the results from MTT assay.

In vivo performance of the nano-reactor
Finally, in vivo behavior of the nano-reactor was explored by using MDA-MB-231 tumor-bearing mice. By virtue of the intrinsic fluorescence of Ce6, the bio-distribution was visualized using a living imaging system. PTFCG@MH or free Ce6 was intravenously injected when the tumor volume reached ~ 100 mm 3 . At 1 h post-injection, Ce6 showed red fluorescence throughout the body (Fig. 5A), indicating non-specific distribution, while most of the fluorescence was cleared after 24 h, with major accumulation in liver. For comparison, the PTFCG@MH displayed much weaker fluorescence at 1 h post-injection due to the fluorescence quenching of the nanosystem, further demonstrating its advantage for decreased phototoxicity. However, the signal at tumor tissue became intensified after 24 h circulation (Fig. 5A, black circle), indicating the EPR effect of the nanomedicine for passive accumulation into tumor. We then quantified the results by collecting the main organs as well as tumor tissues for ex vivo fluorescence imaging (Fig. 5B). PTFCG@ MH exhibited ~ 3.7-fold higher intensity at tumor site than free Ce6, confirming the targetability of the nanosystem towards tumor tissue. Meanwhile, we also noticed considerable accumulation of nanoparticles in liver and kidneys, the main organs to sequester and eliminate nanoparticles from body, and it was also observed in many other nanomedicines [41,42]. Fortunately, temporal controlled activation of PDT would enable selective damage toward tumor, thus minimizing the unwanted side-effects.
To evaluate the in vivo tumor ablation activity, the tumor bearing mice were randomly divided into five groups (n = 5), each intravenous injection of PBS, Ce6 + Laser, PTFCG + Laser, PTFCG@ MH, and PTFCG@MH + Laser, respectively (100 μL, [Ce6] = 2.5 mg/kg), with totally two injections (day 0 and day 4). The laser was performed at 24 h after injection. The therapeutic efficacy was monitored by measuring the tumor size every other day (Fig. 5C). Compared with PBS control, free Ce6 only showed marginal tumor growth inhibition, mainly due to its rapid clearance from mouse body with minimal accumulation into tumor tissue (Fig. 5A). While the nanoparticles could facilitate tumor targeting delivery of Ce6, the efficacy of PTFCG plus laser irradiation was still poor, due to the multiple resistance mechanisms of tumor against PDT therapy. For PTFCG@MH, by contrast, the growth of tumors was obviously suppressed with a strong growth-inhibitory activity (~ 76%). Since we only performed two dosages and the treatment was stopped at day 4, a slight tumor recurrence was observed at day 6. However, the overall tumor growth was significantly suppressed for PTFCG@ MH plus laser as compared with other treatment groups, demonstrating the superiority of the nanosystem for tumor therapy. Note that without laser irradiation, the therapeutic effect of PTFCG@MH was also rather limited, verifying that the anti-tumor activity was mainly originated from the PDT effect.
Next, the tumor tissues were extracted for a series of characterizations. Based on the tumor weight evaluations (Fig. 5D), it was clearly seen that PTFCG@MH plus laser achieved the best efficacy, in consistent with the in vivo measurement. We also explored the pathological changes by H&E staining (Fig. 5E), in which the tumor with PTFCG@MH plus laser treatment displayed widened interstitial space, nuclear condensation, and large vacuoles, indicating the highest level of tumor apoptosis. We also evaluated the key pathological feature of tumor hypoxia by immunofluorescence staining of HIF-1α. The solid tumor showed a relatively high-level HIF-1α expression due to its hypoxia microenvironment (Fig. 5F), and the immunofluorescence was even brighter after PTFCG treatment (with laser) because of the oxygen consumption by GOx catalysis and PDT. For comparison, the tumor hypoxia was effectively relieved after PTFCG@MH (plus laser) treatment, due to oxygenation by MnO 2 . From quantified results, the HIF-1α level was decreased ~ 70% after PTFCG@MH treatment (Fig. 5G). Therefore, the PTFCG@MH could effectively modulate tumor hypoxia microenvironment and provide O 2 substrate to improve PDT treatment outcome.
Finally, the biosafety of the nanosystem was examined. No obvious decrease of body weight was observed over the period of treatments (Additional file 1: Figure S9), and the H&E staining showed no pathological change of the major organs after treatments (Additional file 1: Figure S10). These results indicate the biocompatibility of the nanosystem for in vivo applications.

Conclusions
In conclusion, an intelligent cyclic nano-reactor, PTFCG@MH, was fabricated for enhanced PDT against solid tumor. The nanostructure was well-characterized, and the cyclic reactions were explicitly demonstrated in solution, including GOx-catalyzed glucose consumption, MnO 2 -mediated oxygenation and GSH depletion. The surface HA modification endowed the nano-reactor with improved colloidal stability and active targetability to facilitate its accumulation into tumor after intravenous injection. After being delivered into tumor cells, PTFCG@MH boosted the PDT efficacy via simultaneous ATP/GSH suppression and self-oxygen supply, resulting in efficient tumor growth inhibition with no obvious side-effects. This cascade nano-reactor would promote the development of multifunctional nanoplatforms for improved cancer treatment by modulating unfavorable cancer microenvironment.