Bioorthogonal chemistry has made significant achievements in cell imaging, regulation of biomolecules, and therapeutic applications without interfering with normal bioprocesses. (1) Particularly, it has garnered considerable attention in the field of prodrugs activation for therapeutic treatment of various diseases, especially tumors, which reduces the side effects through in situ drug production. (6,7) Transition-metal catalysts (TMCs) have emerged as ideal candidates for catalyzing bioorthogonal reactions with high catalytic efficiency and bioorthogonal selectivity. At first, a range of TMCs, such as Ru and Pd complexes, have been developed for activating proteins and therapeutic drugs in cells. (8,9) However, the homogeneous catalysts are constrained to low stability and poor biocompatibility in physiological environments. To overcome the drawback of homogeneous catalysts, the biocompatible heterogeneous catalysts are designed to achieve intracellular bioorthogonal catalysis for antitumor treatment. (10-13) For example, Bradley et al. have reported the bioorthogonal generation of 5-fluorouracil catalyzed by Pd-functionalized resins for killing cancer cells. (14) Subsequently, Unciti-Broceta and co-workers realized cancer cell-selective activation of antineoplastic drug with the help of exosome camouflaged Pd nanosheet. (15) To expand the applications of bioorthogonal catalysis, our group presented a covalent organic framework (COF)-based iron catalysts to construct bioorthogonal-activated in situ vaccine for immunotherapy. (16) Despite these achievements, the low atomic utilization and the lack of active sites exposed on the surface of TMCs limit their catalytic activity to a great extent. Besides, the reaction rate of TMCs remains far from satisfactory for generating sufficient active agents under complicated physiological conditions. In terms of improving the catalytic and therapeutic indices of TMCs, the engineering of efficient catalytic platforms with maximum atom utilization efficiency and favorable biocompatibility is highly desirable but challenging.

Single-atom catalysts (SACs) with the atomic-scale active site have made a breakthrough in the field of catalysis. (17-18) As a brand-new class of heterogeneous catalysts, it possesses enhanced catalytic activity and maximum usage efficiency of metallic catalysts, which can bridge the gap between homogeneous and heterogeneous catalysts for various applications. (19-21) In addition, thanks to the active valence electrons and low-coordination environments of metal centers, SACs are emerging as versatile biocatalysts and display superior catalytic activity for various reactions in living conditions. (22-25) It performs prominent catalytic ability in catalyzing various chemical reactions such as oxidation–reduction reactions, photocatalytic reactions, and lactone hydrolysis reactions, which make them fascinating candidates for potentially establishing high-efficient biocatalysts. (26-33) Inspired by these unique properties, we envision that the development of single-atom TMCs with high atom utilization and rich active sites might improve the efficacy of bioorthogonal catalysis for regulating physiological processes.

Herein, for the first time, we developed the sulfur-doped Fe single-atom catalyst as a potent bioorthogonal catalyst for intracellular catalysis and manipulating bioprocesses (Scheme 1). Owing to the atomic-scale active site and high atomic utilization, the Fe-SA catalysts exhibited excellent catalytic activity in the biological milieu even at low concentrations. As a comparative study, the catalytic activity of COF with a similar structure was 1 order of magnitude lower than as-prepared single-atom catalysts under the same conditions. To demonstrate the practicability of catalysts, N6-methyladensoine (m6A) methylation in macrophages is successfully regulated in situ by the designed Fe-SA@Man. m6A methylation is one of the most abundant internal modifications in mRNAs, whose dysregulated methylation has been demonstrated to affect RNA metabolism and cause various diseases, such as cancer. (34-38) By modified with mannose, Fe-SA@Man NCs could preferentially accumulate in macrophages and activate the piperidine-3-carboxylate (MPCH) prodrug in situ, an agonist of m6A writer METTL3/14 complex protein. (39) The synthesized MPCH could upregulate METTL3/14 complex protein expression and cause hypermethylation of m6A modification, which promoted macrophage polarization to the M1 phenotype for antineoplastic therapy. In addition, the NCs were found to display the oxidase (OXD)-like activity to produce reactive oxygen species (ROS), which further boosted the upregulation of m6A methylation (40) and the polarization of macrophages, achieving enhanced antitumor immunotherapy. Using a single-atom catalytic platform, we successfully achieved spatiotemporal precise regulation of m6A RNA methylation of macrophages through bioorthogonal chemistry, which holds great promise for tumor immunotherapy.

To substantiate our design, Fe single-atom catalysts were first synthesized through a pyrolysis method according to the literature with some modifications. (41) l-cysteine, melamine, and FeCl3 were converted to sulfur-doped Fe-SA catalysts eventually. X-ray diffraction (XRD) analysis validated the crystal structure of as-prepared Fe-SA catalysts. The outcome determined that the diffraction pattern matched well with the literature (41) (Figure S1), suggesting the absence of aggregated metallic form in Fe-SA. The morphology of Fe-SA was a two-dimensional network structure displayed by transmission electron microscopy (TEM; Figure 1a). Besides, no metal nanoparticles or clusters existed in Fe-SA as high-resolution TEM (HRTEM) images displayed (Figure S2). According to the aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) analysis, the obvious and individual bright dots belonged to Fe atoms, demonstrating that the Fe atoms of Fe-SA were atomically dispersed (Figure 1b,c). Moreover, the energy-dispersive spectrum (EDS) mapping images revealed the homogeneous distributions of C, N, S, and Fe elements on the entire Fe-SA (Figure 1d). The inductively coupled plasma optical emission spectrometry (ICP-OES) further confirmed that the loading of the Fe atom was 1.0 wt %. Thereafter, the surface area of Fe-SA NCs was determined to be 303.326 m2 g–1 by Brunauer–Emmett–Teller (BET) measurements (Figure S3). Furthermore, X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) were carried out to explore the accurate structure of Fe atoms in Fe-SA. As displayed in the XANES spectrum of Fe K-edge (Figure 1e), the edge position of Fe-SA revealed that the chemical valence of Fe was around +3 (Fe2O3). The Fourier transformed (FT) EXAFS data as illustrated in Figure 1f showed the major peak at 1.50 ？ was ascribed to the Fe–N bond, and the minor peak belonged to the Fe–C backscattering. (21,42) The investigations agreed with the HAADF-STEM results (Figure 1b) and further proved that Fe was atomically dispersed. The Fe sites were determined to be four-coordinated with N atoms, and the Fe–N length was 1.53 ？ by quantitative EXAFS fitting analysis (Figure 1g,h and Table S3). By contrast to the wavelet transform (WT) plots of Fe foil, Fe2O3, and FePc, the Fe-SA displayed the maximum wave transform (WT) corresponding to the Fe–N bond at 4 ？–1, without any WT signal belonging to Fe–Fe (Figure 1i–l). Clearly, the Fe-SA with four nitrogen species coordination was successfully prepared. Afterward, to target tumor-associated macrophages, Fe-SA NCs were functionalized with mannose (termed Fe-SA@Man), which was validated by FT-IR spectra (Figure S4a). Additionally, the amount of mannose on Fe-SA@Man was quantified by thermogravimetric analysis (TGA) for 3.2% (Figure S4b).

The successful fabrication of Fe-SA@Man encouraged us to study their multiple functions. First, the catalytic performance of Fe-SA@Man NCs was investigated by the deprotection of rhodamine 110 derivative (pro-RH 110) in a vial (Figures 2a and S5). (43) As shown in Figure S6, pro-RH 110 was caged so well that fluorescence at 530 nm disappeared. Next, we explored whether the modification of mannose would have an effect on the catalytic performance of catalysts. Figure S7 proves that the catalytic activity was almost unaffected. Meanwhile, the fluorescence intensity was increased with the increasing concentration of Fe-SA@Man (Figure 2b). Thereafter, the changes of the fluorescence spectrum of pro-RH 110 incubated with NCs were monitored from 0 to 60 min to determine the catalytic efficiency and kinetic process. According to Figure 2c,d, the fluorescence intensity kept rising over time in the presence of Fe-SA@Man NCs, while the fluorescence intensity remained static in the absence of NCs. What is more, the photos of different groups demonstrated similar results that the NCs-treated group displayed apparent green fluorescence (Figure S8). Subsequently, to assess the catalytic performance of Fe-SA@Man NCs, structurally similar COF catalysts (denoted as CFe) were prepared as a control based on our previous study (Figures S9 and S10). (16) In particular, the catalytic efficiency of Fe-SA@Man NCs in a simulated physiological environment was extremely high, superior to that of CFe catalysts (Figure S11). The fluorescence intensity of the Fe-SA@Man-treated group was 11.1 times higher than that of the CFe-treated group at 60 min. Besides, the catalytic experiments were carried out in different physiological buffers to evaluate the catalytic ability. As depicted in Figure S12, the catalytic ability of Fe-SA@Man NCs under different conditions was negligibly affected, even in the presence of serum. Notably, in contrast to the deprotection of pro-dye by CFe in a simulated physiological environment, the groups treated with Fe-SA@Man performed with higher fluorescence intensity (Figure S12). The outcomes not only validated the universal applicability of the designed Fe-SA@Man in biological conditions but also exhibited prominent catalytic activity in vitro simulated biological milieu. Furthermore, the catalysts exhibited good reusability (Figure S13) and the supernatant of the catalysts exhibited no catalytic activity after being dispersed in H2O for 3 d (Figure S14). In general, these data suggested that well-designed Fe-SA@Man possessed remarkable catalytic activity. Then, the OXD-mimetic activity of Fe-SA@Man was investigated by the oxidation of 3,3′,5,5′-tetramethylbenzidine (TMB), the oxidation of which displayed a maximum absorbance at 652 nm. (44-46) Given that the catalytic activities of nanozymes were relative to their concentrations, we assessed the OXD-like activity of Fe-SA@Man at different concentrations. As shown in Figure S15, the oxidation rate of TMB was augmented with the increase of Fe-SA@Man concentration. Besides, the OXD-mimicking catalytic ability of Fe-SA@Man NCs was also associated with pH, temperature, and concentrations of substrates (TMB) (Figures S16–S18). Meanwhile, even after 10 catalytic cycles, the Fe-SA@Man NCs still maintained good OXD-mimetic activity (Figure S19). The remarkable activity and stability of Fe-SA@Man were beneficial to its further biomedical application.

Encouraged by the excellent bioorthogonal catalytic activity and OXD-mimetic activity of the designed Fe-SA@Man NCs in vials, their intracellular catalytic performance was investigated by use of RAW264.7 (mouse macrophages) cells. First, the cytotoxicity of Fe-SA@Man was assessed in different cell types including RAW264.7 and 3T3 (mouse embryonic fibroblasts) cells through the classic methyl thiazolyl tetrazolium (MTT) method. After treating for 24 or 48 h, the Fe-SA@Man exhibited a negligible effect on the cell viability of different cell lines, suggesting the good biocompatibility of Fe-SA@Man at the used concentrations (Figure S20). Then, the targeting ability of Fe-SA@Man to RAW264.7 was evaluated by flow cytometry and confocal laser scanning microscopy (CLSM). Figure S21 manifests that Fe-SA@Man could be efficaciously internalized by RAW264.7 cells. Specially, RAW264.7 cells could be reshaped by IL-4 transforming from the resting state (M0) to the M2 phenotype. (47-49) After treatment by IL-4, RAW264.7 cells displayed stronger fluorescent signals in the Cy3-Fe-SA@Man-treated group, superior to the Fe-SA group, suggesting the admirable targeting capacity of Fe-SA@Man to the M2-like RAW264.7 cells (Figure S22). After that, the bioorthogonal deprotection reaction of caged fluorophores catalyzed by NCs was carried out in RAW264.7 cells (Figure 3a). The CLSM imaging showed that stronger green fluorescence existed in the group treated with Fe-SA@Man, suggesting the prominent catalytic activity of Fe-SA@Man catalysts in cells (Figure 3b,c). Meanwhile, a similar result was also obtained by flow cytometric analysis based on the significant fluorescence in macrophages treated with Fe-SA@Man and pro-RH 110 (Figure 3d,e). In addition, a comparison with intracellular catalytic performance was made between CFe and Fe-SA@Man. The data of flow cytometric analysis and mean fluorescence intensity (MFI) showed that the Fe-SA@Man-treated group exhibited a strong fluorescent signal compared with the CFe-treated group (Figure S23). These results definitely manifested that Fe-SA@Man NCs possessed highly efficient bioorthogonal catalytic ability in vitro. Afterward, the 2, 7-dichlorodihydrofluorescein diacetate (DCFH-DA) probe was used to explore the intracellular ROS which was generated by the OXD-like activity of Fe-SA@Man. Once the DCFH-DA probe entered the cells and encountered the oxidative stress, it would be oxidized to the strong fluorescent product 2′, 7′-dichlorofluorescein (DCF). As shown in Figure S24, cells treated with Fe-SA@Man showed gradually bright green fluorescence with the increase of Fe-SA@Man concentration, indicating the occurrence of intracellular oxidative stress caused by its great OXD-like activity. Collectively, Fe-SA@Man could effectively target the M2-like phenotype of RAW264.7 cells and efficiently mediate the activation of pro-fluorescence and the generation of ROS.

The favorable targeting and pro-fluorescence activation ability motivated us to further investigate the intracellular prodrug activation performance of Fe-SA@Man NCs for modulating m6A RNA methylation. In view of the pivotal role of methyltransferases in RNA methylation modification, we chose piperidine-3-carboxylate (MPCH, 4), an agonist of m6A writer METTL3/14 complex protein, as a model prodrug. (39) The catalytic transformation of an inert piperidine-3-carboxylate (pro-MPCH, 3) prodrug into an active MPCH in cells using the constructed catalysts was studied. Preliminarily, the intracellular generation of MPCH mediated by Fe-SA@Man was determined by mass spectrometry (MS). After incubation with Fe-SA@Man + pro-MPCH for 24 h, the cell lysate was collected and the cleavage product, MPCH, was observed (Figure S25). Next, we explored whether the activation of MPCH by a bioorthogonal catalytic strategy could modulate the m6A methylation in macrophages. Above all, after being exposed to pro-MPCH or MPCH with or without Fe-SA@Man NCs as well as sodium ascorbate, the cell viability of RAW264.7 cells did not change obviously, implying the biocompatibility of the bioorthogonal activation strategy (Figure S26). Based on RNA dot blot analysis, cells treated with pro-MPCH alone failed to enhance m6A methylation significantly, indicating that the activity of MPCH was completely inhibited (Figure 4a,b). Meanwhile, the caged active molecule had negligible side effects on the healthy cells (Figure S26). Interestingly, the m6A methylation of cells treated with Fe-SA@Man was substantially improved compared with the control group and prodrug group. These results suggested that ROS played a role in upregulating methylation effect on the m6A. (40) Of special note the highest levels of m6A methylation have been observed in RAW264.7 cells treated with pro-MPCH + NCs, superior to those treated with uncaged MPCH as well. All of these results supported that the activity of pro-MPCH could be restored by bioorthogonal catalysis and cooperate with the generation of ROS to upregulate m6A RNA methylation.

Having verified the effective regulation of m6A methylation by the successful activation of pro-MPCH, we further evaluated their modulating effect on macrophage phenotype. According to previous studies, overexpressing METTL3 promoted polarization of macrophages to antitumor M1-type. (50-52) Therefore, we utilized different formulations to incubate cells and then explored the expression of macrophage markers of different phenotypes by flow cytometry analysis. Figure 4d illustrates that the group with pro-MPCH treatment had no effect on the cell phenotype regulation compared with the IL-4-treated group (group A, M2Φ). A comparison was made between M2 macrophages exposed to Fe-SA@Man alone and those exposed only to IL-4. An increased expression of CD86 (M1 macrophage maker) and a decreased expression of CD206 (M2 macrophage maker) were observed in the Fe-SA@Man-treated group. This could be attributed to the regulation of m6A methylation by generated ROS, which promoted macrophage phenotype conversion. Importantly, in response to pro-MPCH treatment combined with Fe-SA@Man, the generation of M1 macrophage markers was significantly improved. Moreover, the expression level of secreted cytokines including M2-related genes (Arg I and IL-10) and M1-related genes (iNOS and TNF-α) was further detected by qPCR assay. As depicted in Figure 4c, the secreted levels of Arg I and IL-10 were downregulated in group E incubated with Fe-SA@Man + pro-MPCH, while the secreted levels of iNOS and TNF-α were upregulated, suggesting the enhanced secretion of proinflammatory cytokines caused by the in situ-generated MPCH. These results suggested that with the help of Fe-SA@Man NCs, the in situ activation of MPCH cooperated with the generation of ROS could efficiently repolarize macrophages from M2-type to M1-type. Additionally, the phagocytic function of macrophages was explored using a neutral red reagent. (47,53,54) Compared with other treatments, a strong absorption at 540 nm could be detected in group E (treated with Fe-SA@Man + pro-MPCH), giving evidence that the phagocytosis of RAW264.7 cells was enhanced (Figure S27). Together, the regulation of m6A methylation by the bioorthogonal activation of pro-MPCH and generated ROS could effectively re-educate TAMs to M1-type, emphasizing the promise of Fe-SA@Man NCs for accomplishing macrophage-mediated antitumor activity.

Next, we investigated the tumor cell-killing ability of reprogrammed macrophages by a 0.4-μm-sized transwell system. Figure 4e depicts that the RAW264.7 treated with different formulations were incubated in the upper compartment and 4T1 (mouse breast cancer cells) were cultured in the bottom chamber. First of all, the cell-killing ability was evaluated by cell live/dead double staining. A strong red signal could be observed in group E (M2 incubated with Fe-SA@Man and pro-MPCH) compared with other groups, suggesting their outstanding cancer cell-killing ability (Figure 4f). Besides, the MTT assay obtained the same outcomes (Figure S28). Further, the type of cell toxicity was accessed by applying an Annexin V/PI apoptosis assay kit (Figure S29). The results confirmed that apoptosis was the main form of cell toxicity after 4T1 cells cocultured with repolarization macrophage. To sum up, Fe-SA@Man NCs could effectively activate pro-MPCH and generate ROS, promoting macrophage repolarization from M2 to M1 by upregulating m6A methylation, further achieving tumor cell killing.

The outstanding intracellular performance of Fe-SA@Man urged us to access its application in vivo. As a prerequisite, the in vivo short- and long-term toxicology of Fe-SA@Man should be considered. First, the determination of the hemolysis rate was carried out by adding Fe-SA@Man to an erythrocyte suspension (Figure S30). Compared with the water group, the hemolysis rate from 20 to 200 μg mL–1 Fe-SA@Man was less than 1%, complying with ISO 10993–4 standards. (55) It proved that the injection of Fe-SA@Man was available. In the following, the biochemical and hematological analyses were carried out to examine the short- and long-term biosafety, and no apparent toxicity could be observed (Figure S31). Meanwhile, the weight of mice was increased (Figure S31B). After treatment with Fe-SA@Man for 28 days, the mice were sacrificed and major organs were collected for hematoxylin and eosin (H&E) staining. Tissue damage was hardly observed in the Fe-SA@Man-treated group compared with the healthy group, implying the high biocompatibility of Fe-SA@Man (Figure S31C). Besides, the injected sodium ascorbate had no effect on the health of the mice (Figure S32a). Subsequently, the biodistribution of the nanoparticles was discovered by fluorescence imaging. After intravenous injection with Cy3-labeled Fe-SA@Man for 24 h, the major organs were collected for further imaging. As displayed in Figures S33 and S34, masses of Fe-SA@Man were still retained at the tumor site while less of Fe-SA was enriched in tumor regions at 24 h postinjection. These could be attributed to the enhanced permeability and retention (EPR) effect and potent targeting ability of mannose modified on Fe-SA@Man NCs. Overall, Fe-SA@Man exhibited prominent biocompatibility and effective tumor aggregation, which holds great promise for further antiproliferative application.

Subsequently, the antitumor efficacy of Fe-SA@Man was studied by bioorthogonal-activated macrophage reprogramming in a mouse breast cancer model. As a premise, pro-MPCH was found to be stable in the blood of mice (Figure S35). Then, Figure 5a displays the establishment of an orthotopic 4T1 tumor model. The tumor growth curves depicted that the tumor grew rapidly after intravenous administration of pro-MPCH (Figure 5b–d), indicating that the caged-agonist pro-MPCH had little inhibitory effect on tumor growth. Distinguished from the PBS group A, the Fe-SA@Man-treated group C demonstrated better therapeutic effects on tumors. This result might be ascribed to the ROS generated from the oxidase-like activity of Fe-SA@Man that reprogrammed macrophages into the antitumor M1 type by upregulating m6A methylation. Notably, the tumor growth in group E (pro-MPCH + Fe-SA@Man) was significantly suppressed compared with that in group D (treated with MPCH alone), and sodium ascorbate did not affect tumor growth (Figure S32b). We interpreted this phenomenon as the poor efficacy of individual agonists, while the presence of Fe-SA@Man could generate ROS and amplify the methylation level, thus enhancing the therapeutic effect. The H&E staining outcome was consistent with the above results (Figure 5g). Besides, no apparent weight fluctuation of mice was obtained after different treatments (Figure 5e). Meanwhile, the major organs in each group were hardly visible damaged (Figure S36). Moreover, the survival rate of mice was counted and shown in a survival curve (Figure 5f). There was a good correlation between the survival time of each group and tumor inhibition results. As exhibited in Figure 5f, the survival time of mice from group E treated with pro-MPCH + Fe-SA@Man was greatly prolonged, while only 60% of mice survived in the MPCH-treated group after 60 days. Collectively, the in situ regulation of m6A methylation and re-education of TAM mediated by the bioorthogonal catalytic Fe-SA@Man NCs showed remarkable treatment effects with negligible side effects.

To determine the polarization of tumor macrophages in vivo, the 4T1 tumor-bearing mice were sacrificed, and the tumor tissues were harvested for immunofluorescence staining (CD86 and CD206, Figure 6a–d). As shown in Figure 6c, the CD86 expression of tumors in group B treated with pro-MPCH was the same as that in group A treated with PBS, in which hardly a red fluorescence signal was monitored. This result revealed that the caged MPCH would not play a role in repolarizing macrophages. And the Fe-SA@Man-treated group exhibited a slightly enhanced expression of CD86, which was ascribed to the upregulation of the m6A methylation level caused by the OXD-like property of Fe-SA@Man. Strikingly, a remarkable red fluorescence could be observed from group E treated with Fe-SA@Man + pro-MPCH, suggesting the activation of MPCH in situ and the increasing levels of ROS synergistically facilitated m6A methylation upregulation and induced macrophages activation. MFI analysis of immunofluorescence staining revealed that the expression of CD86 in group E was 12.6-fold higher than that in group A (Figure 6d). The infiltration of T cells within the tumor site is interrelated with a good prognosis in cancer treatment. To investigate whether cytotoxic T lymphocytes (CD8+ T cells) and helper T lymphocytes (CD4+ T cells) could play an important role in killing cancer cells and regulating adaptive immunity, the spleens of mice were collected to measure the percentage of T cells (Figure 6e–h). Flow cytometry analysis results displayed that the proportion of CD8+ and CD4+ T cells in group E treated with pro-MPCH + Fe-SA@Man significantly increased (Figure 6e,g). Among them, the percentage of CD3+CD4+ and CD3+CD8+ T cells in Fe-SA@Man + pro-MPCH-treated group were 1.33 and 1.94 times higher than that of the control group, respectively, indicating that CD8+ and CD4+ T cells could be elicited under the synergetic treatment of pro-MPCH and Fe-SA@Man (Figure 6f,h). Above all, our single-atom NCs could immunomodulate the TME by catalyzing the in situ MPCH activation, triggering a powerful immune response.