Assembly of ionizable protein nanocages
Human recombinant ferritin heavy chains can spontaneously assemble into protein nanocages, each composed of 24 heavy chain subunits. This is achievable in E. Coli, using a recombinant protein expression strategy22, 23, 24, 25. In the present study, we utilized a genetically engineered approach to incorporate an ionizable peptide into the N-terminus of the ferritin heavy chain, which results in the display of the ionizable peptide on the outer surface of FTn, yielding iFTn (Fig. 1A). Specifically, the ionizable peptide consisted of nine repeated Histidine-Histidine-Glutamic acid sequences (i.e., 9H2E) (Fig. 1B), capable of ionization under acidic pH conditions due to histidine’s approximate pKa of 6.0, thereby endowing a mechanism facilitating the endosomal escape of iFTn26. Protein structure prediction using AlphaFold2 indicated that the fusion of 9H2E to the N-terminus of FTn resulted in a linear disordered structure displayed on the surface of iFTn, which suggested there would be no interference with folding of FTn subunits. However, the incorporation of 9H2E led to a decrease in the cellular uptake of FTn in H9C2 cardiomyocytes. Among five iFTn constructs (Fig. S1A), iFTn containing 30% H2E-FTn subunits was screened as the optimal protein nanocages to achieve a balance between an ionizable property (Fig. S1B) and satisfactory cellular uptake (Fig. S1C). This iFTn was selected for the subsequent studies, unless otherwise indicated. We then sought to synthesize iFTn into various iFTn-Ner assemblies by using two-armed PEG crosslinkers - a methodology we had previously optimized21. Briefly, after reacting two-armed PEG- succinimidyl carboxylmethyl ester (SCM) with free primary amine (-NH2) groups of FTn, various assemblies were obtained through gel purification. Transmission electron microscopy (TEM) images showed different lengths of various chain-like assemblies (Fig. 1C), including iFTn-1er (11.0 nm), iFTn-2er (20.1 nm), iFTn-3er (29.7 nm) and iFTn-4er (40.6 nm). Different molecular sizes of the four assemblies were further confirmed by native-PAGE gel analysis (Fig. 1D). Size exclusion chromatography analysis also illustrated the larger molecular sizes of various assemblies compared to iFTn and FTn (Fig. 1E). Importantly, the various assemblies exhibited a gradual improvement in ionizable ability with the increase of incorporated iFTn, as evidenced by 6.2 mV for iFTn-1er, 7.1 mV for iFTn-2er, 8.8 mV for iFTn-3er and 9.5 mV for iFTn-4er at pH 5.0, respectively (Fig. 1F). In contrast, positive charges of various iFTn assemblies were not observed at pH 7.2, remaining stable within the range of -10 ~ -20 mV.
Subsequently, we sought to compare the intracellular behavior of different iFTn assemblies. To do this, iFTn was pre-labeled with Cy5 and then assembled with non-labeled iFTn to form various iFTn assemblies through a two-armed PEG, resulting in an equal molar amount of Cy5-labeled iFTn in different assemblies (Fig. 2A). After incubation with H9C2 cells, confocal images showed the distribution of various iFTn assemblies within the cells. Flow cytometry analysis illustrated an increase in signal intensity in cells corresponding to the elevated number of iFTn assemblies (Fig. 2B). Compared to iFTn and iFTn-2er, the uptake of iFTn-4er exhibited an approximately 2.0- and 1.2-fold improvement, respectively. Following uptake, we further explored whether various iFTn assemblies were able to escape from endo-lysosomes. The various Cy5-labled iFTn assemblies were incubated with cardiomyocytes for 2 hours, and following the removal of excess, cells were cultured for an additional 2 hours before fixation and confocal microscopy visualization of Cy5 and LysoTracker stained lysosomes. The images showed the co-localization of the different iFTn assemblies and lysosomes (Fig. 2C). Compared to native FTn alone, enhanced endo-lysosomal escape capabilities of the various iFTn assemblies were observed. Further quantification analysis of the colocalization using the Pearson’s coefficient demonstrated that the iFTn-4er localized in lysosomes decreased by 22% and 18% compared to iFTn-1er and iFTn-2er, respectively. These results revealed the ionizable properties of various iFTn assemblies facilitated their escape from endo-lysosomes, and this phenomenon was dependent on the acquisition of positively charged iFTn particles within the endo-lysosomal microenvironment (i.e., pH 5.0-5.5).
Preparation and in vitro protective effect of iFTn nanozymes
The excessive production of superoxide radicals (O2·-) plays essential roles in driving intracellular oxidative stress. Nanozymes with SOD-mimicking activities are effective in scavenging cytotoxic O2·-. Recently, we employed FTn as a protein scaffold for designing O2·--scavenging nanozymes due to its ability to actively bind with metal ions7. Thus, MnO2 nanozymes were in situ incorporated into the cavity of iFTn (i.e., iFTn-M), using a biomineralization approach, and then were further assembled to form various chain-like iFTn-M assemblies using two-armed PEG crosslinkers, yielding iFTn-M1, iFTn-M2, iFTn-M3 and iFTn-M4 (Fig. 3A). Unstained native-PAGE analysis revealed the visualization of brown nanozyme particles, and the protein shell was shown at the same location after staining with Coomassie Blue (Fig. 3B). Importantly, the electrophoretic migration rates of various iFTn-M assemblies were consistent with their corresponding iFTn assemblies. TEM imaging confirmed the structure of chain-like nanozyme assemblies with a 6.5 nm diameter MnO2 core (Fig. 3C). We next determined whether iFTn-M assemblies possessed SOD-mimicking activity, catalyzing the conversion of O2·- into O2. The SOD-like activities of various iFTn-M assemblies with identical molar concentrations were evaluated in physiological buffer (i.e., pH 7.4). As shown in Fig. 3D, the SOD-like activity of iFTn-M4 is higher than other nanozymes, implying that iFTn-M4 is more effective in scavenging O2·- from the oxidative stress.
Next, we investigated the protective effect of iFTn-M nanozymes against intracellular oxidative stress in cardiomyocytes. Following various treatments, the total level of ROS in H9C2 cells was analyzed using a ROS probe, dichloro-dihydro-fluorescein diacetate (DCFH-DA). Confocal images and quantification analysis by flow cytometry revealed a marked reduction in fluorescence signal intensity after treatment with iFTn-M nanozymes, especially in the iFTn-M4 treatment group, indicating that iFTn-M4 exhibited a significantly greater ROS scavenging ability compared to iFTn-M1 and iFTn-M2 (Fig. S2). Dihydroethidium (DHE), an O2·- probe, can intercalate within DNA for nucleus staining following its oxidization in cells. Confocal images showed bright red fluorescence in the H9C2 cells under oxidative stress, whereas untreated cells did not exhibit fluorescence (Fig. 3E). Under the same oxidative stress conditions, the addition of iFTn-M nanozymes effectively scavenged O2·-, with iFTn-M4 exhibiting the superior scavenging ability. Flow cytometry analysis illustrated that there were approximately 50% and 29% reductions in O2·- following treatment with iFTn-M4 and iFTn-M2, respectively (Fig. 3F). Furthermore, we explored whether these nanozymes protected the cells from oxidative damage. The mitochondrial function was first evaluated by analyzing mitochondrial DNA (mtDNA) copy number via RT-qPCR. There was an elevated protective effect on mtDNA in the iFTn-M4 treatment group, compared to that of iFTn-M2 and iFTn-M1 (Fig. 3G). Cell apoptosis induced by oxidative stress was also analyzed with or without nanozyme treatments. We found that oxidative damage induced approximately 33% of cells to undergo apoptosis, and the addition of iFTn-M nanozymes attenuated apoptosis (Fig. 3H). In particular, the apoptosis percentage of cells induced by oxidative stress significantly decreased by 2-fold following iFTn-M4 treatment (P < 0.0001).
Accumulation and cardiac repair of iFTn nanozymes in mouse IR model
Considering the burst release of O2·- as an initiating factor for cardiac tissue IR injury, we next sought to study whether iFTn-M nanozymes could enhance cardiac tissue repair in a mouse IR myocardial injury model. First, we explored the in vivo biodistribution of various Cy5.5-labeled iFTn using an IVIS imaging system after intravenous (i.v.) administration during IR injury. At 12 hours and 24 hours, ex vivo imaging illustrated that various iFTn assemblies were primarily distributed in liver, with no significant differences among the different assemblies (Fig. S3). Importantly, a significant accumulation of various iFTn assemblies were observed in the isolated IR hearts following i.v. administration (Fig. 4A). Quantitative analysis demonstrated that iFTn-4er had slightly higher accumulation than iFTn-2er and iFTn-1er (Fig. 4B). Assessment at 24 hours indicated that the accumulation difference between iFTn-4er and other iFTn assemblies was larger, with a 2-fold stronger signal intensity in the iFTn-4er group compared to iFTn-2er group (P < 0.0001). These results suggest that iFTn-4er exhibited enhanced accumulation and retention ability in the IR heart, and this was expected, primarily due to the longer blood circulation time of iFTn-4er that we previously reported 21. To further understand their accumulation in the IR heart, we compared the tissue distribution of various iFTn assemblies in sham and IR hearts. After injecting Cy5-labeled iFTn, the hearts were harvested 24 hours later for cryosection, followed by blood vessel immunostaining and manual reconstruction of the cardiac tissue panorama. As previously reported 27, IR induced microvascular damage, leading to reduced vascular density in the heart and noticeable left ventricular enlargement (Fig. 4C), indicating the successful generation of IR injured heart models. In the sham hearts, the signal from various Cy5-iFTn was negligible. In contrast, the signal from various Cy5-iFTn in the injured area was significantly enhanced compared to the sham heart tissue. In particular, in the same IR cardiac tissue, various Cy5-iFTn primarily accumulated in injured area but not in the non-injury area. Consistent with the results of ex vivo imaging, iFTn-4er exhibited the highest particle distribution in IR cardiac tissue. These findings suggest that the improved accumulation ability of iFTn in IR-damage tissue may be attributed to enhanced vascular permeability in injured vessels and elevated expression of FTn receptor in ischemic damaged tissue18, 28.
Next, we assessed the therapeutic efficacy of iFTn-M after three interspaced doses of PBS, iFTn-M1, iFTn-M2, or iFTn-M4 were given through i.v. administration in IR mice (Fig. 4D). At 28 days post-IR surgery, we conducted echocardiography (ECHO) and histological evaluations to assess heart functions. The M-model ECHO images depicted the motion of the left ventricular wall in sham mice and IR mice subjected to different treatments, indicating varying levels of contraction functional damage in the anterior/posterior walls (Fig. 4E). Corresponding quantitative analysis revealed that the parameters of left ventricular ejection fraction (LVEF) and left ventricular fractional shortening (LVFS) in the PBS treatment group had reduced to approximately 50% of the sham mice, as expected (Fig. 4F). Compared to PBS treatment, a gradual improvement in the recovery of heart functionality was observed when mice were administered iFTn-M1, iFTn-M2, and iFTn-M4, as evidenced by the improved parameters of LVEF, LVFS, left ventricular end-diastolic diameter (LVEDd) and left ventricular end systolic dimension (LVEDs). Subsequently, the hearts were isolated for further evaluation of tissue fibrosis using Masson’s trichrome staining. We observed markedly decreased collagen deposition in iFTn-M4-treated mice compared with the levels in the PBS, iFTn-M1, and iFTn-M2 group (Fig. 4G). Further quantification analysis demonstrated that more than 25% of the scar area of the whole heart was damaged for the mice treated with PBS, whereas iFTn-M1, iFTn-M2 and iFTn-M4 treatments decreased to 20%, 15% and 9%, respectively. Based on these findings, the optimized tetrameric iFTn assemblies were pursued for the subsequent construction of cascade nanozymes.
Construction of iFTn cascade nanozymes
Enzyme cascade reactions facilitate the improvement of free radical conversion efficacy. SOD catalyzes the conversion of O2·- to O2 and H2O2, and H2O2 is further converted into non-toxic O2 and H2O by CAT. Thus, an ideal nanozyme should be highly efficient in catalyzing the conversion of O2·- into non-toxic O2 and H2O by possessing both SOD- and CAT-like activities. Inspired by this concept and the benefits of iFTn tetrameric nanozymes (i.e., iFTn-M4), we further constructed an iFTn tetrameric cascade nanozyme by incorporating the high SOD-like activity of iFTn-M and the high CAT-like activity of iFTn-Ru (iFTn-R) (Fig. 5A). Specifically, the cascade nanozyme was assembled by the same molar ratio of iFTn-M and iFTn-R through a two-armed PEG. Theoretically, a typical assembly contained four iFTn nanozymes consisting of two iFTn-M and two iFTn-R (iFTn-MR4), allowing for cascade conversion of O2·- into non-toxic O2 and H2O. Tetrameric iFTn-Mn nanozyme (iFTn-M4) and tetrameric iFTn-Ru nanozymes (iFTn-R4) were also constructed as control nanozymes. The SOD- and CAT-like activities of iFTn-M4, iFTn-R4, and iFTn-MR4 were evaluated using Michaelis-Menten kinetics, a model of enzymatic dynamics. As expected, the SOD-like activity of iFTn-R4 was negligible (Fig. 5B), while its CAT-like activity was significantly higher than iFTn-M4 under the same reaction conditions (Fig. 5C). Importantly, the iFTn-MR4 combined the advantages of the two types of nanozymes, exhibiting both high SOD- and CAT-like activities. The detailed Km (Michaelis constant, the substrate concentration at which the initial rate is half of the maximum velocity) and kcat (turnover number, the catalytic constant for the conversion of substrate to product), two important parameters of enzyme catalytic efficiency, are listed in Fig. 5D.
It is well known that IR injury can induce mitochondrial dysfunction due to ROS overproduction29. We next assessed the protective potential of nanozymes against IR damage, including intracellular oxygen levels, mitochondrial permeability transition pore (mPTP) and mitochondrial DNA (mtDNA) copy number (Fig. 5E). The [Ru (dpp)3]Cl2 is a viable oxygen probe, and its luminescence is strongly quenched by molecular oxygen30. Using this probe, we evaluated the intracellular O2 levels under oxidative stress in H9C2 cardiomyocytes. Confocal images illustrated a significant elevation of oxygen levels after treatment with iFTn-MR4 compared to PBS, iFTn-M4, and iFTn-R4, as evidenced by reduced signal intensity of the probe (Fig. 5F). Further quantitative analysis by flow cytometry confirmed that iFTn-MR4 significantly improved intracellular oxygen levels (P < 0.0001), whereas iFTn-M4 and iFTn-R4 exhibited limited effects (Fig. 5G). Calcium overload leads to the opening of mPTP in the mitochondrial inner membrane, especially when accompanied by oxidative stress31. Confocal images revealed that iFTn-MR4 facilitated alleviation of the opening of mPTP under oxidative stress, as evidenced by the detection with Calcein AM (Fig. 5H). Furthermore, RT-PCR analysis of mitochondrial DNA (mtDNA) copy number was conducted to assess mitochondrial function. The elevated protective effect of iFTn-MR4 on mitochondrial function was significantly better than iFTn-M and iFTn-R (Fig. 5I). These results suggest that the iFTn-MR4 efficiently eliminates intracellular ROS by both SOD- and CAT-like activities, thereby protecting mitochondrial functions in cardiomyocytes subjected to oxidative stress.
In vivo cardiac protection of iFTn cascade nanozymes in IR mouse model
We next investigated whether iFTn-MR4 cascade nanozymes would provide a synergistic protective effect against cardiac IR injury. First, an acute protective action was explored by collecting hearts 24 hours after i.v. administration of nanozymes. The cardiac tissue damage was assessed by determining the observed coloration of six pieces of tissues isolated from each heart using triphenyl tetrazolium chloride (TTC) staining, where the infarct size was determined by white-stained areas (Fig. 6A). The results showed obvious infarct areas within the heart tissues subjected to IR, particularly in the PBS-treated hearts (Fig. 6B). Quantitative analysis of the infarct size in each slice demonstrated a significant reduction in the infarcted area of IR tissue treated with different iFTn nanozymes compared to PBS-treated IR tissue. Importantly, iFTn-MR4 exhibited a synergistic protective effect against cardiac IR injury compared to iFTn-M4 or iFTn-R4 alone.
To further assess long-term protective effects of iFTn-MR4 on cardiac IR injury, we divided IR mice into four groups for three doses of treatments every two days via tail vein injection: PBS, iFTn-M4, iFTn-R4 and iFTn-MR4 (Fig. 6C). After 28 days post-surgery, the structure and function of hearts were evaluated by M-model ECHO imaging (Fig. 6D). Assessment of left ventricle functional parameters, including LVEF, LVFS, LVIDs and LVIDd, demonstrated a gradual improvement in the recovery of cardiac functionality after treatment with iFTn-M4, iFTn-R4 and iFTn-MR4 (Fig. 6E). Notably, various parameters of cardiac functions in the iFTn-MR4-treated group were comparable to that in the sham group. Subsequently, histological assessment of tissue fibrosis was performed using Masson’s trichrome staining (Fig. 6F). Compared to PBS, iFTn-M4 and iFTn-R4, iFTn-MR4 exhibited increased interstitial and perivascular fibrosis. Quantification analysis of the scar area demonstrated that approximately 26% of the total heart area was fibrosed in mice treated with PBS, whereas iFTn-M4, iFTn-R4, and iFTn-MR4 were effective in reducing the total fibrosed area to 14%, 15% and 10%, respectively. The protective effect of the nanozymes on cardiac hypertrophy was also evaluated by Wheat germ agglutinin (WGA) staining. Confocal imaging of cross-sectional area in IR tissue illustrated marked enlarged cardiomyocytes (green) compared to sham tissue (Fig. 6G). In comparison with the PBS group, all IR cardiac tissue treated with nanozymes displayed recovery of cardiomyocyte hypertrophy, with iFTn-MR4 showing the maximal recovery status. Additionally, H&E staining of major organs suggested that there was no apparent accumulation toxicities or damage in the evaluated organs (Fig. S4).
Mechanisms of iFTn cascade nanozymes for mitigating tissue damage
To gain deeper insights into the mechanisms underlying the protective effect of iFTn cascade nanozymes against IR injury, we conducted RNA sequencing (RNA-Seq) analysis to elucidate the gene expression profiles in mouse hearts subjected to myocardial IR injury. The gene expression analysis was performed at 3 days post-surgery following two doses of nanozyme treatments with PBS, iFTn-M4, iFTn-R4, or iFTn-MR4. Venn Diagrams demonstrated that 85% of differentially expressed genes in iFTn-M4 (vs. PBS) and 72% in iFTn-R4 (vs. PBS) overlapped with those in iFTn-MR4 (vs. PBS) (Fig. 7A), indicating that iFTn-MR4 induces changes in gene expression that combine patterns observed in iFTn-M4 and iFTn-R4. The distinct gene expression signatures among the iFTn-M4 (vs. PBS), iFTn-R4 (vs. PBS) and iFTn-MR4 (vs. PBS) are depicted in Fig. 7B, with significant up-regulated genes (red) and down-regulated genes (blue). Using these differentially expressed genes, a Gene Ontology (GO) enrichment analysis was conducted to assess the potential functional improvements conferred by iFTn-MR4. As shown in Fig. 7C, fifteen GO terms related to tissue repair are enriched in iFTn-MR4-treated hearts compared to PBS, mainly including extracellular matrix organization, blood vessel regeneration and remodeling, inflammatory responses, and cellular response to hypoxia. We next analyzed the expression patterns by clustering all differentially expressed genes in the RNA-Seq data. The heatmap image illustrated the alteration of expressing patterns of these genes in PBS-, iFTn-M4-, iFTn-R4-, and iFTn-MR4-treated IR hearts (Fig. 7D). In comparison to the PBS group, iFTn-MR4 induced opposite gene expression patterns, implying functional recovery of IR hearts following iFTn-MR4 treatment. In contrast, iFTn-M4 or iFTn-R4 alone caused changes in only a subset of genes that were altered in iFTn-MR4-treated IR hearts. Further analysis revealed that these differentially expressed genes were associated with superoxide anion, hydrogen peroxide and mitochondrial functions (Fig. 7E). Importantly, iFTn-MR4 induced substantial changes in all listed three aspects of genes compared to PBS-treated IR hearts, whereas different levels of changes were observed in these genes with iFTn-M4 or iFTn-R4 alone. For examples, a significant increase/decrease in the expression of genes associated with superoxide anion, hydrogen peroxide and mitochondrial functions, such as Ncf1, Trap1, Sdhb and Atp5a1, was observed in both iFTn-M4 and iFTn-R4 groups compared to the PBS group. This trend was further enhanced in iFTn-MR4-treated hearts subjected to IR injury. These results implied the synergistic role of iFTn-MR4 in protecting IR hearts by combining therapeutic benefits of iFTn-M4 and iFTn-R4 in mediating superoxide anion, hydrogen peroxide and mitochondrial functions.
To further validate the protective mechanisms of the nanozymes, we assessed superoxide anion-, hydrogen peroxide- and mitochondria-associated functions in IR-injured tissue. Following different treatments in IR mice, we analyzed the levels of O2·- and H2O2 in sham and IR tissue, respectively. DHE histological analysis revealed that nanozyme treatments effectively decreased O2·- levels in IR tissue. Compared to the 80 DHE+ cells per field in PBS-treated IR cardiac tissue, DHE+ cells of iFTn-M4, iFTn-R4 and iFTn-MR4 were decreased to 28 and 8, respectively (Fig. 7F). Quantification analysis of isolated IR cardiac tissue illustrated that, compared to other treatments, the iFTn-MR4 exhibited the lowest H2O2 levels, indicating that the cascade reaction effectively removed H2O2 from the IR tissue, as expected (Fig. 7G). The protective effect of iFTn nanozymes on mitochondrial functions was further analyzed. RT-qPCR analysis of mtDNA copy number within sham and IR cardiac tissues demonstrated significantly elevated mitochondrial functional recovery following iFTn-MR4 treatment compared to PBS, iFTn-M4, and iFTn-R4 (Fig. 7H). Western blot analysis of mitochondrial succinate dehydrogenase complex iron sulfur subunit B (SDHB), a critical enzyme reflecting mitochondrial functions, confirmed that SDHB expression level in IR tissue was substantially recovered after the nanozyme treatments, especially with iFTn-MR4 (Fig. 7I).