Nanofactory for metabolic and chemodynamic therapy: pro-tumor lactate trapping and anti-tumor ROS transition

Graphical Abstract Lactate plays a critical role in tumorigenesis, invasion and metastasis. Exhausting lactate in tumors holds great promise for the reversal of the immunosuppressive tumor microenvironment (TME). Herein, we report on a “lactate treatment plant” (i.e., nanofactory) that can dynamically trap pro-tumor lactate and in situ transformation into anti-tumor cytotoxic reactive oxygen species (ROS) for a synergistic chemodynamic and metabolic therapy. To this end, lactate oxidase (LOX) was nano-packaged by cationic polyethyleneimine (PEI), assisted by a necessary amount of copper ions (PLNPCu). As a reservoir of LOX, the tailored system can actively trap lactate through the cationic PEI component to promote lactate degradation by two-fold efficiency. More importantly, the byproducts of lactate degradation, hydrogen peroxide (H2O2), can be transformed into anti-tumor ROS catalyzing by copper ions, mediating an immunogenic cell death (ICD). With the remission of immunosuppressive TME, ICD process effectively initiated the positive immune response in 4T1 tumor model (88% tumor inhibition). This work provides a novel strategy that rationally integrates metabolic therapy and chemodynamic therapy (CDT) for combating tumors. Supplementary Information The online version contains supplementary material available at 10.1186/s12951-021-01169-9.


Background
The tumor microenvironment (TME) is critically important during the initiation and progression of carcinogenesis [1]. Lactate, has long been considered as the waste product of tumor aerobic glycolysis (termed as the "Warburg effect"), is now emerged as an important regulator of tumor development [2]. Increasing evidence has demonstrated that lactate acted as accomplices to assist tumor immune escape and promote tumor progression, especially through suppressing the function of multiple immune cells, such as macrophages, dendritic cells (DCs), effector T cells [3][4][5]. Therefore, to against cancer, several agents (glycolysis inhibitors (Dichloroacetate, DCA), MCT-inhibiting siRNA (siMCT-4) et al.) have been delivered via nanocarriers to inhibit lactate production and transportation [6,7]. Lactate oxidase (LOX), catalyzing the oxidation of lactate to pyruvate, has also been harnessed for lactate degradation and TME regulation [8][9][10].
The LOX delivery based on nanocarrier represents a promising approach, as it addresses the bio-stability and side effect challenges [8]. Unfortunately, it is still facing the dilemma that the inefficiency of lactate depletion in complex TME and tumor inhibition. Most current works focused on the decomposing lactate with LOX, combined with biozym (glucose oxidase, GOx) or photodynamic therapy to enhance the anti-tumor effect, in which the introduction of additional therapeutic cargo complicated the design of nanomedicine [11,12]. From a different perspective, we propose a strategy of "turn waste into wealth", developed to not only exhaust the lactate but also transform the hydrogen peroxide (H 2 O 2 ), the byproduct of lactate degradation, into cytotoxic drugs via a catalytic medicine approach.
Accordingly, herein, we developed the "lactate treatment plant" (i.e., nanofactory) that can dynamically trap pro-tumor lactate and in situ transformation into antitumor cytotoxic ROS for a synergistic chemodynamic and metabolic therapy. As shown in Scheme 1, cationic polyethyleneimine (PEI), possessing robust acid trapping ability, was chosen as the carrier for loading LOX and Cu 2+ through electrostatic interaction and coordination, named as LNP Cu . The pH-responsive detachable polyethylene glycol (OHC-PEG-CHO) shell endows LNP Cu with stability and biocompatibility (PLNP Cu ). Upon acidtriggered shell shedding, lactate gathered surrounding the "nanofactory" through the "active trapping" of the exposed primary amine on PEI. Then the LOX contained in "nanofactory" facilitated the lactate catalysis to generate H 2 O 2 which was continuously transformed into toxic ·OH through the Fenton-like reaction. In this fashion, the lactate depletion by the cascade catalysis process reversed lactate-induced suppression of immune cells function in TME, and more importantly, its product of ROS subsequently induced ICD and anti-tumor immunity. Therapeutically, the "lactate treatment plant" significantly inhibited the growth of 4T1 breast cancer.

Synthesis and characterization of PLNP Cu
The LNP Cu can quickly assemble by mixing polyethyleneimine (PEI), copper ions (Cu 2+ ) and lactate oxidase (LOX) under gentle stirring at room temperature (see supporting information for the detailed description of the preparation). The driving force for the assembly was mainly electrostatic interaction and coordination among Cu 2+ , PEI and LOX. The obtained LNP Cu have the hydrodynamic size of 115 ± 5.0 nm with a narrow polydispersity index (PDI) of 0.14 by dynamic light scattering (DLS) (Additional file 1: Fig. S1d).
To improve the circulation stability and biocompatibility, the surface of the LNP Cu was shielded with polyethylene glycol derivates with the double end group of aldehyde (OHC-PEG-CHO) through pH-responsive Schiff base (named as PLNP Cu ). As shown in the 1 H NMR spectra, the disappearance (at pH 7.4) and reappearance Scheme 1 Schematic illustration of the lactate exhaustion process of PLNP Cu nanosystem through extracellular lactate active-adsorption and establishment of the intracellular lactate treatment plant (at pH 6.5) of the hydrogen peak for aldehyde groups in the dotted area suggested that the acid-responsive labiality of Schiff bases formed by PEI with aldehyde groups of OHC-PEG-CHO (Additional file 1: Fig. S4). PLNP Cu showed well-dispersed uniform spherical structures (Fig. 1a) with a visible PEG shell at pH 7.4 by transmission electron microscopy (TEM) (Fig. 1b). The hydrodynamic size of PLNP Cu was 218 ± 3.0 nm (PDI 0.06), which was consistent well with the TEM results (Fig. 1d). Notably, PLNP Cu exhibited the compression of zeta potential comparing with LNP Cu (27.07 mV vs. 18.07 mV) (Fig. 1e).
However, the rebound of zeta potential from 18.07 mV to 27.50 mV (Fig. 1e) occurred at the weakly acidic condition (at pH 6.5, the pH of most solid tumor environment) [19], indicating PEG shell could be shed from the surface of PLNP Cu to exposure of the positive charged amino. More intuitively, the shellless PLNP Cu upon incubated in H 2 O at pH 6.5 could be viewed in TEM images, as was shown in Fig. 1c and Additional file 1: Fig. S1a.
Further, the size of PLNP Cu was stable in neutral PBS solution within 48-hour monitoring (Fig. 1f ) accompanied by a low release of LOX (26.25% ± 2.676) ( Fig. 1 h), g LOX loading capacities on LNP and LNP Cu respectively (n = 3). h Cumulative release of LOX from PLNP Cu in PBS at pH 7.4 and pH 6.5 (n = 3). i Hemolysis rate of PLNP Cu at different PEI concentrations (n = 3). Results were expressed as mean ± SD. The significant difference was calculated via two-tailed t-test analysis (g). (NS represented not significant, *p < 0.05, **p < 0.01, ***p < 0.001) the content of LOX was quantified with the Bradford Protein Assay Kit according to the standard curve (Additional file 1: Fig. S3), while the moderate LOX release behavior was found at pH 6.5 (44.74% ± 4.456) (Fig. 1h). The morphology of the PLNP Cu at pH 5.5 was also explored (Additional file 1: Fig. S1b), which suggested the potential for effective function exertion in the intracellular milieu. Apparently, the introduction of copper ion (with entrapment efficiency of ∼81.58% determined by inductively coupled plasma mass spectrometry (ICP-MS)) may facilitate LOX loading in LNP Cu compared with the LNP with a smaller size (Additional file 1: Fig.  S1c) which could be attributed to the coordination of metal ions (LOX %: 226.6 ± 0.518 vs. 219.7 ± 1.237 µg mg −1 ) (Fig. 1g). In addition, hemolytic toxicity tests showed the biosafety of PLNP Cu through intravenous administration (Fig. 1i).

Evaluation of PLNP Cu as lactate treatment plant in vitro
PLNP Cu , referred to as extra/intracellular lactate treatment plant, for lactate consumption and tumor cell inhibition were investigated. As shown in Scheme 2, the lactate treatment plant elicited a chain of the response of processing lactate to antipersonnel reactive oxygen species (ROS) capitalized on the tumor cell physiology and tumor microenvironments (TME), such as acidic microenvironment and the relatively high GSH/GSSG ratios [20,21]. Briefly, the entire process can be categorized into three stages. First, the local high lactate concentration is formed by PEI trapping; second, lactate reacts with oxygen to produce hydrogen peroxide (H 2 O 2 ) under the catalysis of LOX; finally, the Fenton-like reaction between copper ions and enzymatic product (H 2 O 2 ) from the second step produces a large number of hydroxyl radicals (·OH). The verification for the above three stages was investigated as follows.
Inspired by the cationic polymer of PEI with a large number of primary amino groups, we speculated that these positive charged groups could electrostatically interact with lactate and attract to the PLNP Cu nanosystem. We thus carefully studied the adsorption behavior of PLNP Cu . Considering the LOX may influence the determination of lactate absorption of PLNP Cu , bovine serum albumin (BSA) was used to replace LOX, regarding their high similarity in structure and properties (isoelectric point, pI 4.7 vs. 4.6, molecular weight ≈ 66 kDa vs. 80 kDa) except for catalysis for lactate, and was fabricated according to the synthetic procedure of LNP Cu and PLNP Cu (named as BNP Cu and PBNP Cu , respectively). Both zeta potential and hydrodynamic size of BNP Cu and PBNP Cu were consistent with PLNP Cu (Additional file 1: Fig. S2b). Subsequently, the PEI, BNP Cu , and PBNP Cu at different pH with an equimolar amount of PEI were mixed with NaL respectively and shaken overnight at 37 ℃. Followed by fully dialyzed for 4 h and the dialysate of different groups was collected for lactate analysis. The high lactate adsorption rate of BNP Cu (19.54% ± 0.559) and PBNP Cu (pH = 6.5) (13.56% ± 0.0284) were consistent with that of PEI solution (17.24% ± 1.07), while the PBNP Cu (4.369% ± 0.0583) and control groups (5.804% ± 0.429) at pH 7.4 showed much lower levels of lactate adsorption compared to above three groups (Fig. 2a). The above results verified that the exposure of PEI on the PLNP Cu nanosystem played an important role in lactate recruitment, and the detachable PEG shell contributed to the shielding effect in the neutral milieu.
To evaluate the lactate consumption-ability of the nanoparticles, LNP Cu and BNP Cu were added into NaL solution respectively. The concentration of lactate in the mixed solution containing LNP Cu decreased with time compared to BNP Cu (Fig. 2b), which demonstrated the maintenance of the catalytic activity of LOX. We further studied the degradation of lactate over a short time frame (10 or 20 min) at different pH. Surprisingly, after being incubated for 20 min, the lactate degradation rate of LNP Cu was almost four-fold higher than the free LOX group at pH 7.4 (26.16% ± 0.102 vs. 6.727% ± 0.0298), even reaching the ~43 times level in 10 min of reaction (18.71% vs. 0.429%). And LNP Cu exhibited an excellent catalytic activity even at pH 5.5 (7.977% ± 0.122) (Fig. 2c). Immediately after, the degradation rate of lactate following 2 h exposure to LNP Cu was studied. The result showed that the catalytic activity of LNP Cu at different pH values (pH 7.4, 6.5, and 5.5) was significantly higher than the respective values of free LOX groups (Fig. 2d). It was speculated that the discrepancy was likely because of the adsorption effect of PEI in favor of shortening the "lactate-hunting" for LOX, and local high concentration of lactate was more favorable for the Scheme 2 Schematic illustration of the PLNP Cu nanofactory closed-loop for lactate consumption and conversion performance of enzymatic hydrolysis reaction. The above results indicated that the lactate adsorption effect of PEI contributed to the enhance of lactate degradation at different pH.
Based on the above exciting results, the third process in the lactate treatment plant that ·OH production by reacting H 2 O 2 with Cu + was explored. The concentration of H 2 O 2 was quantified via the Hydrogen Peroxide Assay Kit according to the standard curve (Additional file 1: Fig. S5). The LOX solution was mixed with NaL solution and thoroughly reacted for 1 h at 37 ℃ and then the H 2 O 2 concentration of the supernatant was detected. As shown in Additional file 1: Fig. S6 Fig. S7). The methylene blue (MB) degradation was used to validate the Fentonlike reaction occurrence. We simulated the process of intracellular high levels of GSH (5-10 mM) induced the conversion of cupric ions (Cu 2+ ) to cuprous form (Cu + ) [20]. Briefly, GSH was added into the supernatant of BNP Cu and LNP Cu solution respectively for Cu + generation, then the H 2 O 2 and NaL solution were added in turn. The appropriate feeding amount of H 2 O 2 , GSH and NaL was identified from references [22,23]. The reduction in absorbance of MB showed that additional lactate evoked a huge elevation of toxic ·OH in the orange group (with lactate addition) compared to light green groups (without lactate addition) (Fig. 2f ). It was proven that the Cu + resulted from GSH reduction can react with H 2 O 2 to produce ·OH, since the ·OH-induced MB indicator degradation caused the change in absorbance [22]. All the results supported a well-functioning lactate treatment plant in vitro.

Intracellular behaviors of PLNP Cu
As breast tumors exhibit immunosuppression and Warburg phenotype (with strong aerobic glycolysis induced lactate buildup), 4T1 cell line was chosen for further study [24]. We first examined the cytotoxicity of PLNP Cu on 4T1 cells. As shown in Fig. 3a, PLNP Cu at both low and high concentrations induced high cytotoxicity compared with the free LOX group. The higher cell viability upon treatment with PLNP Cu (pH = 7.4) suggested the protective role of the PEG shell at low concentrations. Once the acid-sensitive reaction was triggered, LNP Cu and PLNP Cu (pH = 6.5) showed similar high efficacy for killing 4T1 cells (Additional file 1: Fig. S8). Besides, PBNP Cu exhibited relatively low cytotoxicity even at high concentrations (2.5 µg mL −1 ) (Additional file 1: Fig. S9), which demonstrated that the establishment of an intracellular lactate treatment plant played a critical role in the efficient killing of tumor cells. Then, the cellular uptake of LNP Cu was examined by flow cytometry. 4 h of incubation with LNP Cu @FITC appeared to reach the saturation of cellular uptake for comparable fluorescence intensity as that of 3 h, indicating that the optimal incubation time Additional file 1: Fig. S10. Particularly, both LNP Cu and PLNP Cu (pH = 6.5) had similar high cellular uptakes in 4T1 cells compared to free LOX and PEG-shielded nanoparticles (PLNP Cu pH = 7.4) (Fig. 3b). Taken together, the superior cellular uptake of detached-shell PLNP Cu and establishment of the intracellular lactate treatment plant account for the toxicity of PLNP Cu to 4T1 cells.
Subsequently, the lactate consumption effect of PLNP Cu was evaluated at the cellular level. Notably, the PLNP Cu (pH = 6.5) group showed the lowest levels of lactate among the experimental groups (Fig. 3c), suggesting that the lactate adsorption caused by PEI exposure on the surface of PLNP Cu promoted the lactate consumption process, which was consistent with the functions verification results in vitro. It also showed the potent inhibition of intra/extracellular lactate. The ·OH generation induced by PLNP Cu was investigated using 2′,7′-dichlorofluorescein diacetate (DCFH-DA) as a ROS indicator. From the fluorescent microscopy images, 4T1 cells treated with free LOX and PBS showed negligible fluorescence. And the faint fluorescence in PLNP Cu (pH = 7.4) group might be induced by the endocytosis of a small number of PLNP Cu (Fig. 3d). In contrast, stronger green fluorescence was observed in PLNP Cu (pH = 6.5) treated cancer cells (Fig. 3d), indicating the relatively high ·OH levels rather than H 2 O 2 . Since the mean fluorescence intensity (MFI) in PLNP (pH = 6.5) (the detailed information of PLNP were shown in Additional file 1: Fig. S2a) treated 4T1 cells was insufficient as well (Additional file 1: Fig.  S11). Consistent with fluorescent microscopy images, flow cytometry results also demonstrated the critical role in triggering ·OH formation of copper ions in PLNP Cu (Fig. 3e, f ). Overall, PLNP Cu showed the specific function to convert the lactate into the toxic ·OH product through the intracellular Fenton-like reaction, which indicated the successful construction of the intracellular lactate treatment plant.

Immunomodulatory effects on macrophages and ICD induction effects of PLNP Cu in vitro
·OH, a highly toxic type of ROS, causes oxidative damages to lipids, proteins, and DNA which in turn results in the apoptosis of tumor cells [25]. The apoptosis and necrosis induced by PLNP Cu of 4T1 cells were evaluated by annexin V-FITC and propidium iodide (PI) staining assay. The flow cytometry results showed that the percentage of late apoptosis and necrosis in the PLNP Cu (pH = 6.5) group (56.15%) was nearly 2-fold than other groups (Control ~35.32%, LOX ~34.63%, PLNP Cu pH = 7.4-36.16%), demonstrating the build of the lactate treatment plant in tumor cells and the excellent chemodynamic therapy (CDT) performance (Fig. 4a, b).
We further investigated the immunogenic cell death (ICD) level of 4T1 cells induced by the lactate treatment plant. As three representative markers of ICD, the expression of calreticulin (CRT), the efflux of high-mobility group box 1 (HMGB1) from the tumor cell nucleus, and the secretion of adenosine triphosphate (ATP) were evaluated, which were critical for activating the adaptive immune response and anti-tumor T cell immunity [26]. As the results were shown in Fig. 4c and d, the CRT expression on tumor surface after being treated with PLNP Cu (pH = 6.5) increased to 68.57%, which was considerably 1.29-fold higher than that in the control group. However, the positive rate of CRT in the LOX group (~55.40%) was insignificantly increased compared with controls, indicating that intracellular H 2 O 2 was insufficient to trigger ICD. A similar result was reconfirmed in the immunofluorescence (Fig. 4e). Additionally, HMGB1 is normally localized in the cell nucleus, while the weak green fluorescence at cell nuclei in PLNP Cu (pH = 6.5), demonstrating the large quantities of HMGB1 extracellular released compared with the PBS treatment cells (Fig. 4f ). For ATP secretion, PLNP Cu treatment resulted in higher ATP in the cellular supernatant (Fig. 4g). Notably, it was Fenton-like chemistry driven by cuprous ions that were the main reason for ICD, based on the results of the lower ATP release in PLNP (pH = 6.5) group compared with PLNP Cu (pH = 6.5) (Additional file 1: Fig.  S12). The evidence above suggested that the intracellular lactate treatment plant activated by PLNP Cu converted lactate into anti-tumor ·OH, thereby producing considerable effects on ICD induction in tumor cells, paving the way for anti-tumor immune responses.
As one of the components of innate immunity in TME, macrophages are the major immune component of leukocyte infiltration in the tumor [27]. Unfortunately, lactate efflux associated with lactate accumulation and acid microenvironment suppressed its proliferation and inhibited macrophage M1 polarization [3,28]. The markers (CD80, CD206) of RAW 264.7 macrophages were analyzed by flow cytometric analysis to determine the polarization after stimulating by the PLNP Cu . 1 µg mL −1 of LOX was chosen as the safe dose concentration according to the results of the cytotoxicity experiments (Additional file 1: Fig. S13). The results showed that expression of CD80 (M1 macrophages marker) on the cell surface significantly upregulated with the time increase of PLNP Cu stimulation. Especially, it was up to 9.22% after 24 h of stimulation, which was 3-fold higher than the control group (Fig. 5a, c). Again, immunofluorescence images revealed that the PLNP Cu induced significant upregulation of the M1 biomarker CD80. (Green fluorescence signal of FITC labeled CD80, Fig. 5e). However, the RAW 264.7 macrophages displayed an insignificant change in M2-like polarization via analyzing the expression of CD206 (Fig. 5b, d), which may be due to the expression of CD206 both on M0 and M2-macrophage that compromises the alteration of M2 polarization [29]. Therefore, the M0-macrophage was induced by interleukin-4 (IL-4) for generation of the M2 phenotype and used for M2-macrophages in further studies. The results showed that PLNP Cu upregulated expression of CD80 and, notedly, reduced CD206 expression on M2-macrophages (Fig. 5f, g). We also measured the expression of M1 and M2-associated genes via quantitative realtime polymerase chain reaction (qRT-PCR). The results revealed that PLNP Cu exposure significantly induced M1-related markers expression (CD80, TNF-α) while M2-associated markers (CD206, Arg-1) were downregulated (Additional file 1: Fig. S14). Collectively, these results demonstrated that PLNP Cu induced macrophage Fluorescent microscopy image of CRT exposure. Scale bars, 100 μm. f Fluorescent microscopy image of HMGB1 outflow from the nucleus. Scale bars, 100 μm. g The ATP secretion from cancer cells after different treatments. Results were expressed as mean ± SD. The significant difference was calculated via one- way ANOVA analysis (b, d, g). (NS represented not significant, *p < 0.05, **p < 0.01, ***p < 0.001) M1 polarization in vitro through the consumption of extracellular lactate, which potentiated the antitumoral immune response.

Distribution and antitumor effects of PLNP Cu in vivo
PLNP Cu @Cy7 was prepared to investigate the distribution of the PLNP Cu in 4T1-bearing mice. The major organs (i.e., heart, liver, spleen, lung, and kidney) and tumors were collected within 1 h after tail vein injection of agents (Cy7-labeled LOX, PLNP Cu @Cy7 respectively) for ex vivo fluorescence imaging analysis. As shown in Additional file 1: Fig. S15, free LOX indicated a nonspecific distribution across the normal organs other than the tumor. In contrast, PLNP Cu @Cy7 were significantly enriched in tumor tissues, which might be attributed to the enhanced permeability and retention (EPR) effect [30]. And the enrichment of PLNP Cu @Cy7 in the liver was significantly lower than the free LOX, which decreased the immune-related adverse effects due to a non-specific enrichment of LOX in the liver. In conclusion, these findings suggested that PLNP Cu improved tumor targeting efficiency and accumulation, thereby decreasing the non-specific tissue distribution.
The antitumor effect and induction of systemic antitumor immune responses of PLNP Cu were investigated in the 4T1 tumor-bearing mouse model (a weakly immunogenic tumor) (Fig. 6a). After 18 d of administration (five i.v. injections totally), the PLNP Cu group exhibited the most effective tumor inhibition, while tumors showed similarly rapid growth in PBS and LOX groups (Fig. 6b,  c). Specifically, PLNP Cu demonstrated 88% tumor inhibition, while the LOX only achieved 17%, and PLNP showed a moderate inhibitory of 37% (Additional file 1: Fig. S16). The tumor volume indicated that PLNP Cu effectively inhibited tumor growth in vivo. The same tendencies of body weight changes (Fig. 6d) and the no obvious tissue damage in H&E staining supported the safety of the PLNP Cu treatment in vivo (Additional file 1: Fig. S17).
The effects of the lactate treatment plant PLNP Cu established in the tumor treatment were further discussed in more detail. Tumor tissues were collected and the intratumoral lactate content was detected after five treatments with PBS, LOX, PLNP, and PLNP Cu . The PLNP Cu showed the lowest relative level of lactate (the fold of PBS) in the tumor, at only 0.48, which even reduced to half of the amount for the PBS or LOX group. The result of PLNP (0.79) indicated that thorough consumption and transition of intracellular lactate could counteract the active lactate metabolism in tumors, reflecting in the downregulation of intra/extracellular lactate levels (Fig. 6e). In general, PLNP Cu showed more effective lactate exhaustion than LOX or PLNP attributed to the effective tumor accumulation of nanoparticles and the well-functioning lactate treatment plant.
Although these experiments yielded good results, lactate consumption was only the first step of the lactate treatment plant. The generation of ·OH was quantified to validate the plant working properly. The generation of intracellular ROS in the tumor was detected by flow cytometry with the DCFH-DA probe. The MFI of the intracellular ROS in the PLNP Cu group was 3 times higher than that of the PBS. There was no significant difference between the LOX and PLNP groups and, markedly, both were lower than the PLNP Cu group (Fig. 7c, d).
The above results confirmed that the PLNP Cu triggered the massive generation of intracellular ROS in 4T1 tumor cells. We estimated that the ROS was therefore mainly toxic ·OH that has a high affinity for DCFH-DA binding rather than H 2 O 2 [31]. This conclusion was based on the minor amounts of ROS generation in PLNP group, specifically, the third step of the lactate treatment plant, i.e. Fenton-like reaction, was disrupted due to the absence of copper ion. Incidentally, the limited cell uptake and the insufficient toxic ROS generation of free LOX also demonstrated the positive role of the continuous-operated lactate treatment plant in tumor inhibition. The abundant bright red fluorescence was observed in confocal laser scanning microscope (CLSM) images of frozen tumor tissue sections, consistent with the flow cytometry results respectively (Fig. 7e). Together, the lactate treatment plant establishment triggered by PLNP Cu exhibited the most effective lactate exhaustion and tumor inhibition.

Anti-tumor immune activation in vivo
Encouraged by the well-functioning lactate treatment plant in tumor, we subsequently monitored its antitumor immune responses. For this purpose, the typical cell types of cellular immunity (T cells) and innate immunity (macrophages, Tregs) were detected. The percentage of cytotoxic T lymphocytes (CTL, CD3 + CD8 + T cells) infiltration in tumors after PLNP Cu treatment was significantly increased compared with the PBS (4.06-fold higher) and PLNP (2.09-fold higher). There was a nonsignificant CTL proportion difference between LOX and PLNP groups. Although the slight upregulation of CTL was found in LOX (36.2%), PLNP Cu showed the highest CTL proportion (1.30 times of LOX group) (Fig. 7b). The Fig. 7 Immunogenic CDT induced an immune-active TME in vivo. a Immunofluorescence staining of intratumor infiltrating CD8 + T cells after treatments with PBS, LOX, PLNP and PLNP Cu . Scale bars, 100 μm. b The flow cytometry analysis of tumor infiltration of CD3 + CD8 + T cells. c, d The flow cytometry analysis of ROS generation after treatments. e Immunofluorescence staining of intratumor ROS generation. Scale bars, 100 μm. Results were expressed as mean ± SD. The significant difference was calculated via one-way ANOVA analysis (b, c). (NS represented not significant, *p < 0.05, **p < 0.01, ***p < 0.001) He et al. Journal of Nanobiotechnology (2021) 19:426 same results were observed in CLSM images, in which the copious green fluorescent signal at 488 nm confirmed higher CD8 + T cell infiltration in PLNP Cu group (Fig. 7a).
Furthermore, a typical cytokine (interferon-γ, IFN-γ) of immune activation in tumor tissues was detected. As expected, tumor tissues of PLNP Cu group presented the highest level of IFN-γ after treatment (Additional file 1: Fig. S18). As it is well known, lactate production by cancer cells affects M2-like TAM polarization, which was associated with tumor growth, metastasis, and immunosuppression [3,32]. According to tumor infiltration macrophages analysis, there was a significant increase in M1-like macrophages after PLNP Cu treatment (Additional file 1: Fig. S19). In addition, regulatory T cells (Tregs) in TME inhibit effector T cells and the activation and proliferation of CD8 + T cell by consuming IL-2 and releasing perforin and granzyme [33,34]. PLNP Cu significantly lowered the number of tumor-infiltrating Tregs in immunofluorescence staining results (Additional file 1: Fig. S20). All the above results suggested that immune response activation at the tumor site was triggered through the establishment of the "lactate treatment plant" by PLNP Cu . The systemic anti-tumor immune response induced by PLNP Cu was also explored. The spleen, containing abundant immune cells, is the largest immune organ and the immune center of the body in supporting anti-tumor immune response activation [35]. We analyzed the phenotype and quantity of spleen T lymphocytes. As shown in the flow cytometry results, CD8 + T cells significantly increased in PLNP Cu treatment group, the proportion of CD4 + T cells was downregulated (Fig. 8a). Remarkably, PLNP Cu increased the rate of CD8 + /CD4 + T cells from 0.47 of PBS to 0.69 (Fig. 8d). Dendritic cells (DCs) present the tumor antigen to T cells, playing critical roles in pathogen sensing and initiation of anti-tumor immune responses. The DC maturation in draining lymph nodes (DLNs) including tumor DLNs was analyzed via flow cytometry. By qualitative and quantitative assays, the expressions of the DC maturation marker MHC II, CD86, and CD80 were up-regulated, confirming PLNP Cu promotes DC maturation in DLNs (Fig. 8b, c, e and f ). Together, all the immune analysis results suggesting that PLNP Cu activated the immune system for efficient inhibition of tumor cell growth.

Conclusions
In summary, a pH-responsive nanosystem PLNP Cu was constructed with synergistic anti-tumor effects via the combinatorial treatment with both metabolic therapy and immunotherapy. PLNP Cu delicately used the tumor cell phsiology and TME to build the "lactate treatment plant", where the raw materials (H 2 O 2 that the enzymatic products of LOX, cuprous ions from nanosystem) processed into toxic ·OH products through Fenton-like reactions for ICD, and then the immune system was activated for tumor inhibition. Meanwhile, the intra/ extracellular lactate exhaustion remodeled the immunosuppressive TME, inducing TAM polarization to M1 phenotype and arousing the activity of immune cells around the tumor cells. Importantly, the 88% tumor inhibition was achieved with the PLNP Cu treatment alone without the introduction of additional agents.
Additional file 1. Supporting information including materials, methods, additional figures and tables.