An oligomeric semiconducting nanozyme with ultrafast electron transfers alleviates acute brain injury.

Artificial enzymes have attracted wide interest in disease diagnosis and biotechnology due to high stability, easy synthesis, and cost effectiveness. Unfortunately, their catalytic rate is limited to surface electron transfer, affecting the catalytic and biological activity. Here, we report an oligomeric nanozyme (O-NZ) with ultrafast electron transfer, achieving ultrahigh catalytic activity. O-NZ shows electron transfer of 1.8 nanoseconds in internal cores and 1.2 picoseconds between core and ligand molecule, leading to ultrahigh superoxidase dismutase–like and glutathione peroxidase–like activity (comparable with natural enzyme, Michaelis constant = 0.87 millimolars). Excitingly, O-NZ can improve the 1-month survival rate of mice with acute brain trauma from 50 to 90% and promote the recovery of long-term neurocognition. Biochemical experiments show that O-NZ can decrease harmful peroxide and superoxide via in vivo catalytic chain reaction and reduce acute neuroinflammation via nuclear factor erythroid-2 related factor 2–mediated up-regulation of heme oxygenase-1 expression.

INTRODUCTION

Artificial enzymes have attracted wide interest in biomedicine due to high stability, persistent catalytic activity, and versatile enzyme-like selectivity. In particular, during oxidoreductase, superoxide dismutase (SOD)– and glutathione peroxidase (GPx)–like activity is very important because it is closely related to oxidative stress of the biological system. The antioxidative materials of Mn3O4 hold remarkable SOD-like activity with a clearance rate of >80%, higher than other metal oxide nanoparticles, such as Fe3O4, CeO2, Co3O4, and MnO2 (). However, the GPx-like activity of Mn3O4 nanozyme via surface electron transfer is relatively low with maximum reaction velocity of 0.056 mM min−1 for H2O2. The V2O5 with an active center of V═O and V─O─V exhibits prominent GPx-like activity and serves as an antioxidant nanozyme to achieve cytoprotection (), which catalyzes the reaction at the maximal reaction velocity Vmax of ~0.43 mM min−1 for H2O2 with kcat = 0.065 s−1. Nevertheless, V2O5 nanozyme could exhibit a weak SOD-like activity. It was recently reported that a single-atom enzyme Co/PMCS had multienzyme activities and could rapidly obliterate multiple reactive oxygen and nitrogen species (RONS) to alleviate sepsis (). Co/PMCS single-atom nanozyme demonstrated comparable SOD-like activity to Mn3O4 nanozyme and GPx-like activity to V2O5 nanozyme with the kcat of ~0.116 and 0.086 s−1 for H2O2 and glutathione (GSH), respectively. In addition, although melanin nanoparticles (MeNPs) and Prussian blue nanoparticles (PBNPs) could eliminate the multiple RONS, they failed to mimic GPx (, ). Thus, it is interesting to develop nanozymes with ultrafast electron transfer to achieve redox reactions in an ultrashort time.

Traumatic brain injury (TBI) is frequently associated with posttraumatic stress disorder, memory deficits, and chronic neuroinflammation leading to degenerative brain diseases (, ). The pathophysiology of TBI is multifaceted and can be divided roughly into primary and secondary damage phases. Primary brain damage triggers a cascade of molecular and biochemical events leading to long-lasting secondary neuronal and glial damage, including neuroinflammation, brain edema, and delayed neuronal death (). One of the earliest forms of secondary damage in TBI is oxidative stress (). The brain is particularly vulnerable to oxidative stress due to its high rate of oxygen consumption and high content of transition metals and polyunsaturated fatty acids (). Mitochondrial damage can induce leakage of excessive reduced O2, which forms the primary reactive oxygen species (ROS), O2•− (). Meanwhile, excessive activation of N-methyl-d-aspartate receptors and Ca2+ influx can cause overproduction of reactive nitrogen species (RNS), •NO (). •NO can then react with O2•− to generate highly toxic ONOO−, which damages proteins, lipids, and DNA. Therefore, the balance of free radicals, such as O2•− and •NO, in the first stage of TBI is crucial for limiting the impact of secondary injury and achieving successful treatment of TBI.

Artificial enzymes with enzyme-like characteristics (, –) have recently drawn significant research attention as potential biomedicine (–). Here, we report an oligomeric nanozyme (O-NZ), with a semiconductor core active center and surface-active unit, that achieves ultrafast electron transfer. The core of O-NZ has nitrogen-doped graphite-like structure as multienzyme character, and the surface-active groups include amide groups, hydroxy groups, and pyrrolic N as hydrogen donors to eliminate RONS (Fig. 1, A to C). The nanozyme exhibits an ultrafast core electron transfer of 1.8 ns, achieving the strong SOD/GPx-like activities. Meanwhile, O-NZ holds an ultrafast electron transfer of 1.2 ps between core and surface group, leading to high scavenging activity to O2•-, •NO, and ONOO− free radicals in several milliseconds. The nanozyme treatment significantly increases the overall survival rate of mice with severe TBIs from 50 to 90% and also greatly improves neurocognition and memory by nuclear factor erythroid-2 related factor 2 (Nrf2)–mediated up-regulation of heme oxygenase-1 (HO-1).

Fig. 1.

Structural characterization of O-NZ.

(A to C) Schematic diagram of O-NZ with the nitrogen-doped graphite-like core as multienzyme character and the surface-active groups as hydrogen donors. (D) TEM image of O-NZ. (E) MALDI-TOF MS of O-NZ in a wide range with a regular interval of mass/charge ratio (m/z) of 24. a.u., arbitrary units. (F) Calculated isotope pattern using the molecular model of C91H36, corresponding to the measured pattern. (G) Absorption/emission spectra of O-NZ under 390-nm excitation. PL, photoluminescence. (H) Raman and (I) FTIR spectra of O-NZ. (J to L) XPS spectra of O-NZ for N 1s, C 1s, and O 1s.

RESULTS

Structural characterization and ultrafast electron transfer dynamics

We prepared and screened a series of carbogenic nanozymes with different ligands and surface groups (fig. S1). Among them, generated by ethanol and o-phenylenediamine as precursors, O-NZ was selected because of its high activities. As shown in Fig. 1D and fig. S2, O-NZ has an average core diameter of 2.8 nm and a cross-dimension height of 2.2 nm, as determined by transmission electron microscopy (TEM) and atomic force microscopy (AFM) images, respectively, indicating its ultrasmall size. Results from dynamic light scattering (DLS) reveal an average hydrodynamic size of 3.4 nm for O-NZ (fig. S3), which is notably smaller than the generally agreed-upon renal filtration threshold of 5.5 nm (, ). The molecular weight of O-NZ is experimentally measured to be ~1145 Da by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) and calculation of isotope patterns (Fig. 1, E and F, and fig. S4), indicating that the nanozyme is short oligomer. Compared with the other three nanozymes (L-NZ, C-NZ, and G-NZ), the ultraviolet-visible (UV-vis) absorption spectrum of O-NZ shows a specific absorption peak at ~432 nm (Fig. 1G and fig. S5), which can be categorized into the n-π* transition of the surface states containing C═O/C═N and C─O/C─N structures. Similar to the previous reported carbon dots (), the absorption peaks at ~230 to 310 nm are assigned to the π-π* transition of the aromatic C═C bond. A series of emission spectra of O-NZ under different excitation wavelengths reveal two emission bands (fig. S6A). The emission band at short wavelength of O-NZ shows strong excitation dependence, which is dominated by surface defects (). Different from the L-NZ, C-NZ, and G-NZ, O-NZ also has an excitation-independent emission peak at 556 nm originating from direct exciton recombination (fig. S6). Therefore, the energy gap of O-NZ calculated from absorption and emission spectra is ~2.23 eV, which is consistent with the electrochemical result (fig. S7). Raman spectrum of O-NZ exhibits two peaks at around 1640 cm−1 (G band) and 1390 cm−1 (D band) (Fig. 1H) corresponding to the E2g vibration mode of sp2 carbon domains and structural defects or partially disordered structures of sp2 domains (), respectively. The upshift of G band as compared with graphene (1580 cm−1) is associated with nitrogen doping (). Fourier transform infrared (FTIR) spectroscopic spectrum shows N─H or O─H stretching, C═C stretching, amide N─H bending, and C─NH─C stretching bands (Fig. 1I), suggesting the presence of hydroxy and amino groups in O-NZ. This is very different from FTIR spectra of L-NZ, C-NZ, and G-NZ (fig. S8). Moreover, x-ray photoelectron spectroscopy (XPS) analyses of O-NZ indicate the presence of C═C, C═O, and C─O bonds, especially several coexisting nitrogen-related signals, such as C═N, pyridinic N, pyrrolic N, and graphitic N (400.9 eV) (Fig. 1, J to L), similar to reports on chemically modified graphite using nitric acid and other nitrates (–). In addition, N/C radio and O/C radio of O-NZ are enhanced than L-NZ, C-NZ, and G-NZ (table S1), revealing the high N-doping and abundant hydroxy groups of O-NZ. Therefore, O-NZ has nitrogen-doping graphitic structure as its core and amide, hydroxy, and pyrrolic groups as surface functional groups.

To gain further insight into the electron transfer dynamics of O-NZ, we performed fluorescence lifetime and femtosecond transient absorption (TA) measurement. Figure 2A shows the fluorescence decay curve of the O-NZ, and the average decay time is estimated to be 1.8 ns, much smaller than the general carbon quantum dots (6.6 to 8.3 ns) and Au clusters (13.3 ns) (, ). This indicates that the internal core of nanozyme has ultrafast electron transfer. From the femtosecond TA spectra with excitation at 340 nm (Fig. 2B), O-NZ exhibits two strong negative stimulated emission peaks at around 410 and 560 to 620 nm, coinciding with its emission peaks (fig. S6). Moreover, the positive features correspond to excited-state absorption at around 460 nm. Figure 2C shows the kinetics of each featured wavelength in terms of delay time, indicating that one part of the coulomb-induced hot carriers at carbon core is trapped by surface states, including the carbon backbone and functional groups within 1.2 ps. Therefore, O-NZ has ultrafast electron transfer characteristic. As shown in Fig. 2D, energy level diagram of O-NZ is assumed according to the electrochemical method. The calculated highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels are −5.717 and −3.441 eV, respectively (fig. S7). As a result, the electron transfer path of O-NZ was proposed in Fig. 2E. Given the nature of the carbon core, the π-conjugated domains of O-NZ are likely to facilitate charge transfer and electron storage (, ). Besides, the structural defects of O-NZ could promote the interfacial electron transfer (), and surface functional groups of O-NZ can also serve as electron donors (). Therefore, O-NZ exhibits an ultrafast core electron transfer of 1.8 ns, which may determine the enzyme-like properties. Meanwhile, an ultrafast electron transfer of 1.2 ps occurs between core and surface ligand group, which may determine the radical scavenging activities of O-NZ.

Fig. 2.

Ultrafast electron transfer dynamics of O-NZ.

(A) Time-resolved photoluminescence spectrum of O-NZ. (B) TA spectra of O-NZ at indicated delay times from 2 ps to 7.7 ns. Abs, absorbance; mOD, milli-optical density. (C) Kinetic traces at different probe wavelengths. (D and E) Schematic diagram of ultrafast electron transfer process of O-NZ. LUMO, lowest unoccupied molecular orbital; HOMO, highest occupied molecular orbital. ESA, excited-state absorption; SE, stimulated emission.

Superior RONS scavenging activity and enzyme-like activity

Using the 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid (ABTS) assay, we find that O-NZ exhibits strong scavenging activity against ROS in a manner, which was superior to other carbogenic nanozyme analogs (L-NZ, C-NZ, and G-NZ) by approximately an order of magnitude (Fig. 3A), indicating a strong electron-donating ability to ROS (, ). In addition, the scavenging capability of O-NZ shows concentration dependent (Fig. 3B and fig. S9A). Time-dependent absorbance curve reveals the fast reaction rate of O-NZ against ABTS+• (fig. S9B). Notably, fig. S10 shows that O-NZ could effectively and continuously scavenge 1,1-diphenyl-2-picrylhydrazyl (DPPH), an ideal compound composed of several N radicals with unpaired electrons, indicating strong overall scavenging capability against RNS. Furthermore, compared with L-NZ, C-NZ, and G-NZ, O-NZ exhibited a strong enhancement in general RNS scavenging capability (fig. S11). Further, we evaluated whether O-NZ has the enzyme-like activities. The SOD-like activity of O-NZ was tested by the Total Superoxide Dismutase Assay Kit with WST-8 (S0101, Beyotime). As shown in Fig. 3C, the catalytic disproportionation of O2•− to H2O2 and O2 enhances with the increasing concentration of O-NZ. Compared with the reported antioxidative materials such as V2O5, Mn3O4, CeO2, MeNPs, PBNPs, PCNSs (porous carbon nanosphere), and N-PCNSs (fig. S12), O-NZ exhibits the highest SOD-like activity (fig. S13). Besides, the GPx-like activity of O-NZ was measured by the glutathione reductase (GR)–coupled assay. Figure 3D shows that the concentration of reduced form of nicotinamide adenine dinucleotide phosphate (NADPH) decreases with the presence of O-NZ. The descending rate gets faster with the increase in O-NZ, H2O2, and GSH concentrations (Fig. 3D and fig. S14, A and B), indicating the GPx-mimicking activity of O-NZ. Moreover, as shown in fig. S14 (A and B) of the steady-state kinetic analyses, the GPx-like activity of O-NZ follows a Michaelis-Menten kinetics. The Vmax for H2O2 and GSH obtained from the direct analysis using GraphPad Prism software is 72.97 and 18.49 μM min−1, respectively. The calculated Michaelis constant (KM) for H2O2 and GSH is 0.8204 and 0.8688 mM, respectively. Compared with natural GPx enzyme (8.695 mM), the KM of O-NZ for GSH is much lower (fig. S14C), indicating the higher affinity of O-NZ. In addition, the GPx-like activity of O-NZ is much higher than the reported antioxidative materials except V2O5 (fig. S15). Overall, O-NZ has stronger SOD and GPx-mimicking activities and catalytic efficiency due to its ultrafast electron transfer property.

Fig. 3.

Superior and wide-ranging RONS scavenging activity, SOD-like, and GPx-like properties of O-NZ.

(A) Comparison of the total antioxidant capacities of L-NZ, G-NZ, C-NZ, and O-NZ against ROS. (B) Concentration-dependent investigation of ABTS•+ in the presence of O-NZ. (C) Inhibitor rate of SOD-like activity for O-NZ. (D) GPx-like activity of O-NZ monitored as a function of time. (E to H) O2•−, •OH, •NO, and ONOO− scavenging activities of O-NZ, respectively. (I) Illustration of RONS scavenging routes of O-NZ, schematic diagram of the SOD-like and GPx-like activities of O-NZ, and GSH recycling by GR. NAPD+, nicotinamide adenine dinucleotide phosphate.

To investigate the scavenging capabilities of O-NZ in more details, we monitored multiple RONS species individually. For O2•− and •OH scavenging, electron spin resonance (ESR) spectra were performed (Fig. 3, E and F). The ESR signals from the spin adducts significantly decreased with increasing O-NZ concentrations. Notably, the scavenging efficiency of O-NZ for O2•− was significantly superior to that for •OH. In addition, we investigated the •NO and ONOO− scavenging capability of O-NZ. To probe the •NO scavenging capability of O-NZ, S-nitroso-N-acetylpenicillamine (SNAP) was used to serve as the source of •NO. Carboxy-PTIO (2-Phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide) with five-line electron paramagnetic resonance (EPR) signals was used to trap •NO and generate carboxy-PTI (2-Phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl) with seven-line EPR signals (Fig. 3G). In the presence of O-NZ, the ESR signals showed five lines, indicating that the carboxy-PTIO was not reduced into carboxy-PTI by •NO, and, in other words, •NO was completely cleared by O-NZ. This revealed the strong scavenging activity of O-NZ for the highly active •NO essential for TBI treatment. As ONOO− has a strong characteristic absorption peak at ~302 nm, UV-vis spectrophotometry was adopted to evaluate ONOO− scavenging. As shown in Fig. 3H, ONOO− could be completely eliminated by O-NZ even at a relatively low concentration of 11.8 μM. As low as 3 μM O-NZ could still gradually scavenge ONOO− over time (fig. S16). These facts revealed that O-NZ exhibited a profound scavenging ability for ONOO−, important in treating TBI. This is because ONOO− is produced as the intermediate product from •NO and O2•− under oxidative stress and can cause extensive cellular damage following TBI. These results demonstrate that O-NZ reveals high scavenging activity for RONS, especially for more active O2•− and RNS, which can be attributed to the synergistic effect of amide, hydroxy, and pyrrolic surface groups as hydrogen donors and more active centers.

Therefore, a mechanistic scheme of O-NZ is summarized to show the scavenging process against multiple RONS and enzyme-like activities (Fig. 3I). Gao and coworkers () reported that N doping and graphitic structure were critical for carbogenic nanozymes to perform enzyme-like activities and the enzymatic activities were enhanced when N-doping ratio were increased. Moreover, hydroxyl groups as the hydrogen donor or electron donor were important for phenolic antioxidants, such as α-tocopherol and flavonoids, to exhibit high radical scavenging activities (). Besides, the structural defects of O-NZ could promote the interfacial electron transfer between O-NZ and RONS (). In addition, the ESR spectrum indicates that O-NZ in solutions has an intrinsic unpaired electron (fig. S17), which is capable of quenching the RONS. On the basis of the above analysis, we conclude that the synergistic effect of graphitic structure, high N doping, abundant hydroxy groups and defects, as well as ultrafast electron transfer, could account for the excellent RONS scavenging activity and enzyme-like activity.

Inhibition of cellular oxidative stress in vitro

We first tested the cytoxicity of O-NZ on N2a (mouse neuroblastoma N2a cells), MA-c (mouse astrocytes-cerebellar), and primary neuronal cells using the adenosine 5′-triphosphate (ATP) assay, which shows low cytotoxicity (fig. S18, A to C). Next, the O-NZ effect on viability of cells stimulated with H2O2 and lipopolysaccharide (LPS) was investigated. After nanozyme treatment, the survival rate of H2O2-stimulated N2a cells increases from 76% to the level of the untreated control, while in LPS-stimulated N2a cells, it increases from 65 to 89% (fig. S18, D and G). Similar phenomena are observed for MA-c and primary neuronal cells (fig. S18, E, F, H, and I). Moreover, the survival rate increases in a dose-dependent manner in both cases, demonstrating the protective effect of O-NZ on nerve cells.

Further, in vitro ROS scavenging was investigated using various probes of general ROS including O2•−, •OH, peroxynitrite, and •NO (Fig. 4, A to H, and fig. S19). Figure 4 (A and B) and fig. S19 show that intracellular ROS increased 5.7 and 4.4 times in cells treated with H2O2 and LPS compared with normal cells, indicating increased cellular oxidative stress resulting from the stimulus. However, the total ROS level significantly decreases with O-NZ treatment, close to the control level, which is consistent with previous work (, ). Similarly, flow cytometry analyses show that O2•−, •OH, ONOO−, and •NO free radicals could be appreciably reduced with O-NZ treatment (Fig. 4, C to H), suggesting enzyme-mimicking properties of O-NZ to scavenge multiple ROS and RNS under physiological conditions consistent with Fig. 3.

Fig. 4.

O-NZ markedly reduces intracellular ROS, cell cycle arrest, and apoptosis induced by H2O2 in vitro after 24 hours.

(A) Fluorescent images of general ROS detected in H2O2 treated N2a cells using 2′,7′-dichlorofluorescin diacetate (DCFH-DA) probe and (B) corresponding general ROS level by the flow cytometer. (C) Fluorescent images of •OH level in cells using hydroxyphenyl fluorescein (HPF) probe and (D) corresponding quantitative results obtained from flow cytometer. (E) Fluorescent images of O2•− level detected using dihydroethidium (DHE) probe and (F) corresponding quantification of O2•− detected by the flow cytometer. (G) Fluorescent images of •NO level tested by the fluorescent probe 3-amino,4-aminomethyl-2‘,7‘-difluorescein diacetate (DAF-FM DA) and (H) the corresponding quantitative result using flow cytometer. (I) Cell apoptosis of H2O2-induced N2a cells with and without O-NZ treatment and (J) corresponding apoptosis cells analyzed by flow cytometer. (K) Cell cycle analyzed by counting the percentage of DNA after propidium iodide (PI) staining. (L) Cell cycle phase analysis based on the flow cytometer results. Data were presented as means ± SD (n ≥ 3). One-way analysis of variance (ANOVA) was used for statistical analysis. *P < 0.05.

We next examined the effects of O-NZ on function and cell cycle in H2O2- and LPS-damaged neurons (Fig. 4, I to L, and fig. S20). It is found that apoptotic cells account for up to 51% of all cells in H2O2-treated neurons and 59% in LPS-treated neurons (Fig. 4, I and J, fig. S20, A and C). Meanwhile, numerous cells are arrested in the S or G2-M phases based on cell cycle analysis after treatment with H2O2 and LPS (Fig. 4, K and L, and fig. S20 B and D), which indicates damage to DNA structures or key enzymes involved in mitosis, eventually leading to checkpoint failure. Nevertheless, O-NZ greatly relieves ROS- and RNS-induced cell damage. The percentage of apoptosis and ratio of cell cycle arrest are restored close to the normal levels. Oxidative stress induced by H2O2 and LPS at high concentrations can disrupt normal cell function and metabolism, lastly resulting in massive apoptosis, necrosis, and cell cycle arrest (). Nanozymes with multiple scavenging properties are effective in controlling damage induced by RONS bursts and consequently against oxidative stress.

Improvement of survival rate and secondary spread of injury of TBI mice

To further investigate O-NZ as a potential TBI treatment, we randomly divided and subjected C57/BL6 mice to brain injury followed with intravenous injection of O-NZ at 1 hour after injury. Meanwhile, edaravone, a clinically used ROS scavenger, was used as the positive control. To evaluate the spread of secondary damage following an initial TBI, time-dependent pathophysiologic and recovery progresses in vivo were monitored using magnetic resonance imaging (MRI) (Fig. 5A). Hyperintense T2-weighted lesion (red arrow) can be observed in mice 5 days after TBI, indicating edema or dilated ventricles resulting from accumulation of water. Minor hyperintensity is also observed anterior of the damage. However, 17 days after injury, the TBI mice show aggravated edema deep into the white matter with enlarged and enhanced hyperintense lesions (Fig. 5B). In the edaravone-treated group, aggravated edema is still present in the brain at 17 days after injury. In contrast, there is no detectable damage at the injury site in the O-NZ treated group. Thirty days after injury, significant hyperintensity is still observed in the TBI mice (Fig. 5, A and B), while mice that received O-NZ treatment show negligible edema, suggesting the ability of O-NZ to effectively reduce the spread of secondary damage. The survival rate was recorded over 30 days of observation (Fig. 5C). Mice following TBI show high mortality with 50% survival. In the edaravone-treated group, the animal survival rate is increased to 60%, possibly due to inhibition of ROS production. The effect of O-NZ is far superior to edaravone, increasing the survival rate to 90%, which is a tremendous improvement in the high mortality of severe and acute TBI. Evans blue (EB) staining shows high influx of dye with visible hematoma 1 day after TBI (Fig. 5, D and E), suggesting increased blood-brain barrier (BBB) permeability caused by TBI (). Thus, O-NZ can rapidly penetrate across the damaged BBB and be trapped in the injured region for a long time to play the therapeutic role (fig. S21). After 3 days, however, the O-NZ-treated group shows significant recovery of BBB integrity. Moreover, apoptosis staining reveals that mice with TBI have widespread apoptotic cells distributed along the damaged area of the cortex, but O-NZ significantly reduces these apoptosis signals (Fig. 5, F and G). These results indicate that O-NZ treatment could effectively inhibit the spread of secondary damage of TBI mice up to 1 month via early recovery of BBB and vitality of neurocyte.

Fig. 5.

O-NZ reduces the spread of secondary damage after TBI in mice.

(A) T2-weighted MRI was performed on TBI, TBI + edaravone, and TBI + O-NZ groups at the time points of 5, 17, and 30 days after injury. The spread of secondary damage was visualized by hyperintense regions surrounding the primary damage (red arrows). Slices from 1-mm anterior and 1-mm posterior to the impact site were also shown. (B) Quantification of volume of hyperintense secondary damage through all slices from T2-weighted MRI. (C) Survival rate of O-NZ–treated TBI mice within 30 days. Edaravone was set as positive control. (D) Brain images at 1 and 3 days after injury after intravenous injection of EB solution. (E) Quantification of BBB permeability from EB staining of O-NZ–treated and untreated TBI mice at 1 and 3 days (n = 3 per group). (F) Apoptotic population in brain tissue at 3 days after injury using terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) method and (G) corresponding apoptosis quantification (n = 3 per group). Data are presented as means ± SEM. ***P < 0.001 versus the control group and #P < 0.05 and ##P < 0.01 versus the TBI group, analyzed by one-way ANOVA.

To further evaluate that O-NZ treatment could promote neurological recovery after TBI, we subjected mice to a series of behavioral tests. The Morris water maze (MWM) was used to investigate TBI-induced long-term cognitive deficits (). During the acquisition phase of days 14 to 18, 30 to 34, and 60 to 64, all the mice apparently learned the task, while the distance traveled and latency to hidden platform with O-NZ treatment obviously decrease compared with TBI group (Fig. 6, A and D, and fig. S22, A to D). In particular, O-NZ treatment for 1 month is shown to normalize the performance of searching time for platform, suggesting enhanced spatial learning and acquisition ability (Fig. 6D). For the probe trial on days 19 and 35 (Fig. 6, B, C, E, and F), the number of platform crossings and the percentage time in the missing platform quadrant are significantly reduced in the TBI group, while almost returning to the normal level after O-NZ treatment, revealing that O-NZ treatment could efficiently improve the spatial memory of TBI mice. In addition, the O-NZ-treated TBI mice show comparable spatial learning and spatial memory with normal mice even on day 65 (fig. S22, D to F), suggesting the improvement of long-term cognitive capability. Other time-dependent behavioral tests are investigated in Fig. 6 (G to I). In the rotarod performance test for motor function, time spent on the rod increases significantly and reaches the normal level 100 days after TBI in the O-NZ-treated group, as compared to the untreated TBI mice, indicating recovery of motor coordination and physical condition (Fig. 6G). Injured mice start to show neurological deficits within hours after trauma, which aggravates with time and peaks after 3 days, but O-NZ could be able to lower neurological severity score (NSS) to an extent (Fig. 6H), suggesting recovery of neurological functions. Moreover, in the corner test, O-NZ treatment shows a tendency to normalize behavior in 1 month (Fig. 6I). In summary, O-NZ treatment contributes to distinct recovery in terms of NSS, motor function, spatial learning, and memory after 30 days, while untreated mice show persistent deficits.

Fig. 6.

O-NZ treatment causes behavioral improvement in TBI mice.

(A to F) MWM test (n = 10 per group). (A and D) Latency to locate and rest on the hidden platform recorded for spatial learning trials on days 14 to 18 and 30 to 34. (B and E) Number of platform location crossings and (C and F) the percentage of time during the probe trial spent in the quadrant of tank that previously housed the hidden platform recorded for the probe trial on days 19 and 35, evaluating the reference memory of each mouse. Data are presented as means ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 analyzed by one-way ANOVA. (G) Latency to fall from the rod in the rotarod test at different time points after injury (n = 7 per group). O-NZ treatment can improve the motor coordination and balance deficits induced by TBI. (H) NSS assessed at different time points after injury (n = 7 per group). The NSS obtained before injury was set as baseline. (I) Right turn number ratio quantified in the corner test when stimulating both sides of the vibrissae simultaneously (n = 7 per group). Data are presented as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus the control group; and #P < 0.05, ##P < 0.01, ###P < 0.001, and ####P < 0.0001 versus the TBI group, analyzed by one-way ANOVA.

Reduction of inflammatory factors in TBI mice

To probe the mechanisms for O-NZ–mediated improvement of outcome in TBI, we assessed oxidative stress and neuroinflammation in brain tissues. As shown in Fig. 7A, large amounts of NO are produced in the brain 1 and 7 days after TBI, which is an inflammatory product that is implicated in secondary damage. After the O-NZ treatment for 14 days, the NO level nearly returns to normal level. The H2O2 level increases significantly at day 1 in TBI mice (Fig. 7B). In contrast, O-NZ treatment markedly reduces H2O2 back to almost normal brain tissue levels after 14 days. Similarly, there is an increase in the oxidized glutathione (GSSG) and lipid peroxide level in untreated TBI mice compared with the control at day 1, while O-NZ could rescue GSSG and lipid peroxide to the normal level after 14 days (fig. S23). On the basis of the GSH/GSSG ratio, elevated oxidative stress could be significantly detected at day 1 in TBI mice, while it is recovered closed to normal level under O-NZ treatment (Fig. 7C). Meanwhile, SOD activity in O-NZ-treated mice is restored to normal levels 7 days after injury (Fig. 7D), reflecting the significant antioxidative and RONS-scavenging ability of O-NZ.

Fig. 7.

O-NZ treatment reduces oxidative stress and restores healthy levels of inflammatory proteins in the brain of TBI mice.

(A to D) Oxidative stress indicators including NO, H2O2, GSH/GSSG, and SOD in brain tissue of TBI mice with or without treatment of O-NZ at 1, 7, and 14 days after injury (n = 5 per group). (E to G) Immunofluorescence images for astrocytes (GFAP), microglia (Iba-1), and neurons (NeuN) in the injured cortex at 7 days after injury with or without O-NZ treatment, respectively. Blue indicates the 4’,6-diamidino-2-phenylindole (DAPI) staining of cell nuclei. (H) Quantitative analysis of the pixel density of GFAP/Iba-1/NeuN cells in the injured cortex with or without O-NZ treatment (n = 3 per group). (I and J) Enzyme-linked immunosorbent assay (ELISA) quantitative analysis of IL-1β and TNF-α levels in brain tissues on days 1, 3, and 14 with or without O-NZ treatment (n = 5 per group), respectively. Healthy mice injected with PBS or O-NZ were named as control or O-NZ group, respectively. Data are presented as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus the control group; and #P < 0.05, ##P < 0.01, and ###P < 0.001 versus the TBI group, analyzed by one-way ANOVA.

As representative indicators for neuroinflammation induced by TBI (, ), reactive astrocytes and microglia were studied by markers of glial fibrillary acidic protein (GFAP) and Iba-1 (Ionized calcium binding adapter molecule 1), respectively. As shown in Fig. 7E and fig. S25A, astrocytes undergo reactive astrogliosis characterized by morphological and functional adaptations including up-regulation of GFAP in the injured cortex on days 1 and 7 after injury. O-NZ treatment could significantly reduce the number of reactive astrocytes (Fig. 7H). Concurrently, microglia undergo a similar transformation in morphology (Fig. 7F and fig. S24B), and the number of reactive microglia 7 days after injury shows a little rise than that 1 day after injury (Fig. 7H and fig. S24D), which is consistent with the reported literatures (, ). Meanwhile, neurons marked with NeuN (neuronal nuclei) markedly reduce in injured cortex after TBI while it is significantly elevated via O-NZ treatment (Fig. 7G and fig. S24C), indicating the neuroprotective effect of O-NZ. Moreover, two important proinflammatory factors, tumor necrosis factor–α (TNF-α) and interleukin-1β (IL-1β), show gradual restoration 14 days after treatment (Fig. 7, I and J), suggesting recovery of the neuroimmune system.

To further investigate the antioxidation and antineuroinflammatory mechanism of O-NZ, we performed real-time quantitative polymerase chain reaction (RT-qPCR), Western blot experiments, and immunohistochemistry (IHC) staining (Fig. 8). It is well known that Nrf2 has considerable neuroprotective effects in central nervous system (CNS) diseases (). Once stimulated by trauma, the total Nrf2 expression in the perilesional cortex is markedly elevated displayed by IHC staining (Fig. 8, B and C). O-NZ treatment could further increase the number of Nrf2 positive cells to enhance antioxidant response and reduce inflammation. As the downstream proteins of Nrf2 signaling pathway via activating the antioxidant response element, HO-1 expression was evaluated by IHC staining. As shown in Fig. 8 (D and E), HO-1–positive cells increase with the up-regulated expression of Nrf2 in the O-NZ-treated group. Similarly, the mRNA levels of Nrf2 and HO-1 in the cerebral cortex of the injured side by qPCR and Western blot analyses also demonstrate that O-NZ could ameliorate brain oxidative stress and neuroinflammatory by up-regulating Nrf2 and HO-1 expressions (Fig. 8, F to H). In addition, the induction of other antioxidant responsive element–driven genes, NAD(P)H: quinone oxidoreductase-1 (NQO-1), was assessed by qPCR method. The NQO-1 follows the same trend as HO-1 (fig. S25). These comprehensive results from oxidative stress and neuroinflammatory and protein/gene expression manifest the molecular mechanism of O-NZ for TBI repair, that is, O-NZ via multicatalytic processes can gradually balance in vivo RONS and reduce neuroinflammatory by activating the Nrf2/HO-1 signaling pathway, similar to previous work ().

Fig. 8.

Molecular mechanism of O-NZ for brain injury treatment.

(A) Schematic diagram of TBI repair molecular mechanism of O-NZ treatment via Nrf2-mediated up-regulation of HO-1 expression. (B and C) IHC images of Nrf2 and quantitative analysis of Nrf2-positive cells in the perilesional cortex. (D and E) IHC images of HO-1 and quantitative analysis of HO-1–positive cells in the perilesional cortex at 7 days after TBI after O-NZ treatment. (F) The mRNA levels of Nrf2 and HO-1 in the cerebral cortex of the injured side by qPCR analyses at 7 days after TBI after O-NZ treatment. (G and H) Western blot quantitative analysis and Western blot bands of total Nrf2 and HO-1 in the cerebral cortex of the injured side at 7 days after TBI after O-NZ treatment. Data are presented as means ± SEM. **P < 0.01, and ****P < 0.0001 versus the control group; and #P < 0.05 and ####P < 0.0001 versus the TBI group, analyzed by one-way ANOVA (n = 3 per group).

In addition, toxicological effects of O-NZ on TBI mice were tested by body weight, clinical hematologic, and biochemical indicators. It is found that the injection of O-NZ in healthy mice does not lead to weight loss compared with control group and the injection of O-NZ in TBI mice causes weight gain compared with TBI group (fig. S26). The biodistribution of O-NZ in the healthy mice and TBI mice shows the similar trend that kidney owns the high uptakes due to renal clearance, but there is a little difference in the brain owing to the BBB damage (fig. S27). Last, to evaluate the pharmacokinetics, excretion, and long-term in vivo toxicity, we intravenously injected O-NZ into healthy mice. O-NZ also shows acceptable stability in both phosphate-buffered saline (PBS) and fetal bovine serum (FBS) (fig. S28). The nanozymes exhibit a short blood half-life (~26.3 min) and a rapid excretion, with renal clearance reaching 80% 12 hours after injection (fig. S29). The long-term in vivo toxicity assessments (figs. S30 to S32), including hematology, biochemistry, and pathology studies, show that O-NZ does not cause obvious toxic reactions after 3 months, suggesting no long-term adverse effects.

DISCUSSION

TBI causes a series of secondary injuries that last from hours to months after the initial trauma and lead to irreversible neuron loss (, ). TBI-triggered oxidative stress contributes significantly to the pathological progression of brain injury. () Therefore, inhibition of RNS and ROS accumulation at early stages immediately after TBI is essential to mitigate secondary brain damage. Improvement of the survival rate of severe TBI has been proven to be challenging, with very few compounds showing survival benefits clinically (, ). O-NZ, however, was shown in our study to provide an extraordinary improvement in the survival rate, indicating a potentially feasible solution for the high mortality rate of TBI. In addition to the survival benefits, O-NZ–treated TBI mice also presented functional improvements in behavior and normalization of brain inflammation levels. In contrast, untreated TBI mice suffered persistent neuronal damage and behavioral and cognitive difficulties. O-NZ treatment helps to maintain an appropriate balance of the highly active •NO molecule and acts as an enzyme-like molecule to supplement native SOD. Meanwhile, its catalytic activity further allows O-NZ to reduce excessive levels of H2O2 and prevents subsequent ROS reactions and glutathione oxide production. Likely as a result of this RONS-neutralizing activity, the levels of neuroimmune factors such as TNF-α and IL-1β in TBI mice can be reduced to normal levels after receiving O-NZ treatment.

Nanozymes represent an important class of therapeutics for TBI treatment, but only nanozymes with high catalytic selectivity can be potentially useful for CNS diseases. O2•− and •NO are primitive free radicals widely distributed in the brain, serving as key targets for brain-targeting antioxidants. However, most previous work on nanozymes has focused on improving the scavenging efficiency for •OH, while few reports targeted O2•− (, , ) and •NO scavenging (). O-NZ displays its unique advantages on scavenging RNS, such as •NO and ONOO−, thanks to its high catalytic selectivity. Note that •NO was restored to healthy levels 90 days after injury, indicative of in vivo catalytic balance after treatment. Nitric oxide is an important signaling molecule, and if it is not removed thoroughly and timely, then it may trigger new catalytic responses (). Moreover, unlike inorganic nanozymes (–), organic nanozymes such as O-NZ show fast excretion and low toxicity. This preclinical study presented the therapeutic potential of nanozymes in treating TBI, but further investigational work is needed, such as probing the effects on specific cell types and determining intra- and intercellular mechanisms of O-NZ. In the future, enzyme-like molecules with rational design on functionalization may prove interesting for TBI treatment.

In summary, O-NZ, exhibited high RONS scavenging activity and SOD/GPx-like catalytic selectivity via ultrafast electron transfer. O-NZ showed excellent scavenging capacity for O2•− and ONOO−, as well as the highly active •NO. O-NZ improved the survival rate of acute TBI mice to 90%, compared with only 50% of untreated TBI mice. Furthermore, O-NZ improved the long-term behavior of TBI mice, including spatial memory, cognition, and functional limb control, while irreparable injury and permanent cognitive impairment were observed in TBI mice without treatment. Tissue level analyses found that TBI treatment with O-NZ could restore antioxidant enzymes, active signaling molecules, inflammatory factors, and oxidative stress indicators to healthy levels. The molecular mechanism of O-NZ for TBI repair is that O-NZ via multicatalytic processes can gradually balance in vivo RONS and reduce neuroinflammatory by activating the Nrf2/HO-1 signaling pathway. This work clearly manifests that ultrafast nanozymes with both catalytic activity and selectivity are featured as an important therapeutic candidate for treatment of nervous system diseases.

MATERIALS AND METHODS

Materials preparation

Preparation of O-NZ

Two grams of o-phenylenediamine was dissolved in 200 ml of ethanol, and then the reaction solution was thermostated at 180°C for 1 hour under high pressure to obtain a brownish yellow solution. The ethanol was removed by rotary evaporation, and then the resulting product was redissolved in water. The aqueous solution was ultrafiltered through a 3000-Da ultrafiltration tube to collect the lower filtrate. The lower filtrate was dialyzed by a 1000-Da dialysis bag for 3 days. Last, the yellow solution in dialysis bag was collected and dried to powder as the final purified product for further testing and characterization. The concentration of O-NZ was quantified by weight and absorption spectrum.

Preparation of L-NZ

One gram of lysine was dissolved in 50 ml of ultrapure water, and the mixture was transferred to a beaker and carbonized at 90°C for 30 min in a microwave oven. To remove the impurities, the ultrafiltration process was performed as abovementioned. The final solution was dried to powder for further testing and characterization. The concentration of L-NZ used was quantified as abovementioned.

Preparation of C-NZ

One gram of lysine and 0.276 g of cysteine were dissolved in 50 ml of ultrapure water, and the final molar ratio of lysine and cysteine was 3:1. The mixture was transferred to a beaker and carbonized at 90°C for 30 min in a microwave oven. After the abovementioned ultrafiltration process, the final solution was dried to powder for further testing. The concentration of C-NZ used was quantified as abovementioned.

Preparation of G-NZ

One gram of lysine and 0.701 g of GSH were dissolved in 50 ml of ultrapure water, and the final molar ratio of lysine and GSH was 3:1. The mixture was transferred to a beaker and carbonized at 90°C for 30 min in a microwave oven. After the abovementioned ultrafiltration process, the final solution was dried to powder for further testing. The concentration of G-NZ used was quantified as abovementioned.

Materials characterization

TEM imaging was conducted on a JEOL JEM-2100F instrument with an accelerating voltage of 200 kV. AFM image was performed on a MultiMode 8 microscope (Bruker) with tapping mode in air. The hydrodynamic size of O-NZ was measured by DLS method on a Zetasizer Nano ZS (Malvern, UK). Absorption spectra were recorded on Shimadzu 3600 UV-vis–near infrared (NIR) spectrophotometer. Emission spectra were measured on a Jobin Yvon Fluorolog-3-21 fluorescence spectrophotometer under different excitation wavelengths. XPS was operated at 300 W on a 5400 PHI ESCA spectrometer with a monochromatic Al Kα x-ray source. FTIR spectroscopy was conducted on an IRAffinity-1S spectrometer. MALDI-TOF MS was performed on an AB SCIEX TOF MS 5800. The luminescence lifetime was measured by a Fluorolog-3 fluorometer (HORIBA Jobin Yvon Inc., France), using 375 nm as the excitation wavelength.

Transient absorption

Femtosecond pump-probe TA measurements were performed using a regenerative amplified Ti:sapphire laser system (Coherent; 70 fs, 6 mJ per pulse, and 1-kHz repetition rate) as the laser source and a femto-TA100 spectrometer (Time-Tech Spectra). Briefly, the 800-nm output pulse from the regenerative amplifier was split in two parts with a 50% beam splitter. The transmitted part was used to pump an optical parametric amplifier that generated a wavelength-tunable laser pulse from 250 nm to 2.5 μm as pump beam. The reflected 800-nm beam was split again into two parts. One part was attenuated with a neutral-density filter and focused into a 2-mm-thick sapphire or CaF2 window to generate a white light continuum used for probe beam. The probe beam was focused with an Al parabolic reflector onto the sample. After the sample, the probe beam was collimated and then focused into a fiber-coupled spectrometer with complementary metal-oxide semiconductor sensors and detected at a frequency of 1 kHz. The intensity of the pump pulse (340 nm) used in the experiment was controlled by a variable neutral-density filter wheel. The delay between the pump and probe pulses was controlled by a motorized delay stage. The pump pulses were chopped by a synchronized chopper at 500 Hz, and the absorbance change was calculated with two adjacent probe pulses (pump-blocked and pump-unblocked). The samples were placed in 1-mm quartz cuvettes and were vigorously stirred in all the measurements.

Energy level of O-NZ

The conduction band (CB) and valence band (VB) of O-NZ were tested by electrochemical method using a standard three-electrode system. The three electrodes included Ag/Ag+ electrode as the reference electrode, Pt wire as the counter electrode, and glassy carbon electrode as the working electrode. Ammonium hexafluorophosphonate (180 mg) and ferrocene (5 mg) were dissolved into anhydrous acetonitrile (5 ml) to prepare the electrolyte. After bubbling N2 through the electrolyte for 3 min, the cyclic voltammetry (CV) curve of ferrocene was obtained at a sweep rate of 50 mV s−1. The CV curve of O-NZ was obtained by the same experimental method. The redox potentials of ferrocene and O-NZ were acquired by the analyses of CV curves. According to the formulawe can calculate the VB top (EHOMO), CB bottom (ELUMO), and bandgap (Eg) of O-NZ.

RONS scavenging capacities of O-NZ

For total antioxidant capacity against ROS for O-NZ, we measured strictly in accordance with the instructions of T-AOC Assay Kit (S0119, Beyotime) based on the ABTS agent. Briefly, each tube was added in sequence with 20 μl of peroxidase, 10 μl of O-NZ with different concentrations, and 170 μl of ABTS working solution. After a 6-min incubation, the resulting solution was measured by UV-vis spectrophotometer at 405 nm. The time-dependent absorbance at 734 nm was performed in dynamic mode. The L-NZ, C-NZ, and G-NZ with the same molar concentration were used as contrast samples.

For total antioxidant capacity against RNS of O-NZ, 50 μM DPPH• dissolved in a mixture of dimethyl sulfoxide (DMSO) and water (1:40) was used as a highly stable nitrogen-centered radical. The scavenging reaction was initiated with the addition of 10 μl of O-NZ to 190 μl of DPPH, and the absorption spectrum was immediately measured by a UV-vis-NIR spectrophotometer. The time-dependent absorbance at 543 nm was performed in dynamic mode. The L-NZ, C-NZ, and G-NZ with the same molar concentration were used as contrast samples.

For O2•− scavenging of O-NZ, an ESR spectrometer (Bruker EMX plus, Germany) was operated at 0.6-mW microwave power and 1-G modulation amplitude. KO2 and 5-tert-butoxycarbonyl-5-methyl-1-pyrroline N-oxide (BMPO) were used as the contributor and spin trapper of O2•−, respectively. The resulting solution contained 2.5 mM KO2, 25 mM BMPO, and 3.5 mM 18-crown-6 in DMSO medium. ESR spectra were recorded with or without the presence of O-NZ at different concentrations.

For •OH scavenging of O-NZ, we used ESR spectroscopy under the same settings as mentioned above. •OH were originated from 5 mM hydrogen peroxide in 10 mM buffer under UV irradiation for 5 min and trapped by 50 mM BMPO. ESR spectra were recorded with or without the presence of O-NZ at different concentrations.

For ONOO− scavenging of O-NZ, UV-vis spectrophotometry was used in the range of 200 to 500 nm. The ONOO− solution should be freshly prepared. H2O2 (10 ml, 50 mM), HCl (5 ml, 1 M), and NaOH (5 ml, 1.5 mM) were quickly added in that order into NaNO2 (10 ml, 50 mM) with quick stirring in an ice-water bath. The resulting ONOO− solution showed pale yellow. Absorption spectra were immediately measured after the addition of O-NZ at different concentrations. The time-dependent absorption spectra were recorded in real time.

For •NO scavenging of O-NZ, 250 μM SNAP was used as the source of •NO, and ESR spectroscopy was used under the same settings as mentioned above. Carboxy-PTIO (10 μM) with five-line EPR signals was used to trap •NO and generate carboxy-PTI with seven-line EPR signals. ESR spectra were recorded with or without the presence of 10 μM O-NZ.

Enzyme-like activity of O-NZ

The SOD-like activity of O-NZ is determined by the Total Superoxide Dismutase Kit. The steps are as follows: For each reaction, 250 μl of SOD detection buffer, 75 μl of nitro blue tetrazolium (NBT) coloring solution, and 0.75 μl of enzyme solution were mixed to make 400 μl of NBT/enzyme working solution. We added 50 μl of O-NZ, 400 μl of NBT/enzyme working solution, and 50 μl of reaction starter in order. The reaction was performed under light for 20 min, and the absorption value at 560 nm was detected by a spectrophotometer. The blank control group without O-NZ needed light. The light control group without O-NZ did not need light.

The GPx-like activity of O-NZ was studied by GR-coupled analysis method. The spectrophotometry was used to detect kinetics at 340 nm by following decrease in the concentration of NADPH (the molar extinction coefficient is 6.22 mM−1 cm−1). In experiments, the volume of the total reaction solution was fixed at 500 μl, and the order of adding the reactants was O-NZ, GSH, NADPH, GR, and H2O2. The experiments were performed by varying the concentration of H2O2 (0.025 to 1.2 mM) and the concentration of GSH (0.25 to 10 mM) at a time and fixed the concentration of GR (1.7 U) and NADPH (200 μM) in PBS (100 mM) (pH 7.4) at 25°C. Michaelis-Menten curves were drawn by the direct analysis using GraphPad Prism software. Natural GPx enzyme was also measured the enzyme kinetic parameters under the same experimental condition.

Cell toxicity and viability

Mouse neuroblastoma (N2a) and MA-c cell lines were purchased from National Infrastructure of Cell Line Resource. Primary neuronal cells were extracted from the cerebral cortex of mouse embryos in pregnant C57BL/6 mice. Cells were cultured in an incubator containing 5% CO2 at 37°C with Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) containing 10% FBS (Gibco) for N2a and MA-c cells or B-27 Plus Neurobasal medium (Gibco) for primary neuronal cells, respectively.

At a concentration of 5 × 105 cells ml−1, cells were seeded into a 12-well plate for N2a, MA-c, and primary neuronal cells, respectively, and cultured in an incubator until cells were completely adherent. Then, culture medium in each well was replaced by fresh DMEM or Neurobasal medium containing O-NZ at different doses. Cells from the control group were treated without O-NZ, and the wells without cells were regarded as blank. The cytotoxicity was assessed after 24-hour incubation by the ATP assay. According to the manufacturer’s instructions of ATP assay kit (S0026, Beyotime), after centrifugation, cell debris was removed, and the supernatant was added into the substrate solution. The luminescence of solution was recorded in an Illuminometer with an integration time of 1 s per well. The protein content was tested according to the BCA Protein Assay Kit (P0012S, Beyotime), and then the ATP concentration was converted to nanomoles per milligram of protein. ATP concentration was directly proportional to cell viability. For the tests of cell survival rate, cells were inoculated in a 12-well plate for N2a, MA-c, and primary neuronal cells, respectively, and incubated with 200 μM H2O2 (88597, Sigma-Aldrich) or LPS (0.5 mg/ml; L8880, SolarBio) for 6 hours. Then, the culture medium was substituted by fresh medium containing O-NZ at different doses. The plates were kept in the incubator for 24 hours, and cell viability was assessed by the ATP assay.

RONS scavenge ability in vitro

N2a cells (1 × 105 cells per well) were cultured in six-well plates for 24 to 48 hours before treatment with 200 μM H2O2 or LPS (0.5 mg/ml) for 6 hours. Then, the solution was replaced by fresh-culture medium with 10% FBS containing 40 μM O-NZ. The intracellular oxidative stress was measured using various fluorescent probes that target specific RONS, including 2′,7′-dichlorofluorescin diacetate (DCFH-DA) [Ex (excitation) = 488 nm and Em (emission) = 525 nm] (D6883, Sigma-Aldrich) for general ROS, dihydroethidium (DHE) (Ex = 535 nm and Em = 610 nm) (D7008, Sigma-Aldrich) for superoxide, hydroxyphenyl fluorescein (HPF) solution (Ex = 490 nm and Em = 515 nm) (H4290, Sigma-Aldrich) for •OH, and 3-amino,4-aminomethyl-2′,7′-difluorescein diacetate (DAF-FM DA) (Ex = 495 nm and Em = 515 nm) (S0019, Beyotime) for NO level. According to the protocols from the manufacturer, cells were kept at 37°C for the appropriate period of time in the dark. Intracellular oxidative stress was captured using a fluorescence microscope (EVOS, AMG), and quantitative analysis of free radicals was conducted by a fluorescence-activated cell sorting flow cytometer (BD Accuri C6). The quantification were presented with self-fluorescence of O-NZ deducted.

Apoptosis and cycle analysis

Cells that reached complete attachment in six-well plates were treated with 800 μM H2O2 or LPS (2.5 mg/ml) for 6 hours, and then the solution was replaced by fresh culture medium with 10% FBS containing 40 μM O-NZ for 24 hours. Next, cells were washed with PBS three times at 1500 rpm for 10 min. The cell concentration was adjusted to 106 cells ml−1 in PBS or binding buffer for cell cycle or apoptosis analysis, respectively. Apoptosis were detected by incubating 100 μl of cell suspension in sequence with 5 μl of fluorescein isothiocyanate–annexin V for 5 min and 5 μl of propidium iodide (PI) for 10 min under dark. The suspension was then diluted by 300 μl of binding buffer for flow cytometry (BD Accuri C6). For cycle analysis, all cell samples were fixed with 70% cold ethanol drop wise and kept at 4°C for 12 hours. Cells were collected by centrifugation (1500 rpm for 15 min) and mixed with 100 μl of ribonuclease at 37°C for 30 min. Then, PI solution was added to each sample for another 30 min. Cell cycle was analyzed using the flow cytometer (BD Accuri C6).

Traumatic brain injury

Male C57BL/6J mice at 6 to 8 weeks were purchased and maintained under the guidelines approved by the Key Laboratory of Post-trauma Neuro-repair and Regeneration in Central Nervous System, Tianjin Medical University General Hospital. Animals were raised under specific pathogen–free level environment in a regular 12-hour dark/light cycle and supplied with food and water ad libitum. Surgery was performed after 1 week of transportation to adapt to the environment. Mice were anesthetized with 10% chloral hydrate (10 mg/kg) by intraperitoneal injection to conduct fluid percussion injury (FPI) models. The mouse was then placed in a stereotaxic frame, and the scalp was cut to expose the skull. The craniotomy was performed by drilling the skull above the right side of the parietal-temporal cortex in the circle of 4 mm in diameter. A plastic female Luer lock disc was cemented to the craniotomy site using cyanoacrylate adhesive followed by dental cement and was allowed to wait for 15 min for the disc to be firmly fixed over the site. Then, the mouse was positioned under the FPI device (Custom Design & Fabrication, Richmond, VA) such that the affixed disc was in the same plane as the end of the device connected to it. The head of the mouse was held firmly and subjected to a moderate or severe injury pressure (mortality experiments, 2.6 ± 0.1 atm; behavioral assessment experiment, 1.9 ± 0.1 atm). After injury, the Luer lock fitting was gently removed, and the scalp was stitched up. All mice were placed on a warm pad to recover from anesthesia and then divided into control, O-NZ, TBI, TBI + O-NZ, and TBI + edaravone group. Mice were intravenously injected with O-NZ immediately after injury at the dose of 50 mg/kg with the injection volume of 200 μl. Edaravone was used as positive control and injected intravenously at the dose of 4 mg/kg based on body surface area. As control, healthy mice injected with PBS or O-NZ at the dose of 50 mg/kg were named as control or O-NZ group, respectively. Animal survival was recorded during 30 days of observation (n = 20 per group).

Magnetic resonance imaging

Brain tissue in vivo (n = 4 per group) was examined by MRI at 5, 17, 30 days after TBI on a 3-T MR scanner (MR750, GE Healthcare, USA) to generate T2-weighted images. Images were acquired using a fast spin echo pulse sequence: TR (time of repetition)/effective TE (time of Echo) = 3093/77 ms, echo train length = 22, matrix = 256 by 256, field of view = 80 mm by 80 mm, and 12 slices of 1 mm in thickness. Mice were injected with 10% chloralic hydras and put into the MR apparatus. T2-weighted MRI were performed for a group of four mice. Volume of secondary damage was estimated from hyperintense regions indicative of edema in T2-weighted images using ImageJ software.

BBB permeability

EB solution (2%) was intravenously injected into injured mice (n = 3 per group) on days 1 and 3 after injury. Forty minutes later, mice were perfused with PBS three times (10 ml for each time). The brain tissue was taken out and separated into the ipsilateral and contralateral sections. All samples were then homogenized in N,N-dimethylformamide (Sigma-Aldrich) and centrifuged at 9000 rpm for 20 min. The supernatant was then collected for permeability evaluation. EB was quantified by measuring optical absorption at 620 nm.

Apoptosis by terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling assay

Mice (n = 3 per group) were perfused with PBS (20 ml) and paraformaldehyde (10 ml, 4%), respectively, on day 3 after injury, and the brain tissue was gathered and kept fixed for 48 hours. The brains were then dehydrated in gradient ethanol (100, 95, 80, and 70%) before being embedded into paraffin and sliced into 4-μm coronal sections. Tissue apoptosis was tested using terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate (dUTP) nick end labeling (TUNEL) assay kit (C1090, Beyotime). Specifically, tissue sections were deparaffinized in dimethylbenzene and rinsed in gradient ethanol to water. The slides were incubated in proteinase K (20 μg/ml) at 37°C for 20 min. After rinsing in PBS three times, the slides were incubated with terminal deoxynucleotidyl transferase enzyme and Cy3-labeled dUTP mixture and kept at 37°C, avoiding light for 1 hour. Tissue apoptosis was then examined using fluorescence microscope (BD Accuri C6) and quantified using ImageJ software.

Behavior tests

MWM test

To evaluate the spatial learning and spatial memory, MWM test was performed in a circular water pool on days 14 to 19, 30 to 35, and 60 to 65 after TBI (n = 10 per group). A circular stainless steel tank with 122 cm in diameter and 51 cm in height on both sides with nonreflective interior surfaces was used. The water made opaque by milk was maintained at 24° to 26°C with an automatic heater to avoid hypothermia, and the platform was set in the center of quadrant I. Before the spatial learning, visual discrimination learning was performed to distinguish whether the vision of mice was normal. On the five consecutive days for spatial learning trials, each mouse was tested with four trials a day approximately at the same time each day with an intertrial interval of 15 s, and the start location was randomly selected in four different quadrants. Each mouse was put into the water facing toward the tank wall and allowed to swim for 60 s to find the hidden platform below water surface. Mice that failed the task would be placed onto the platform and left there for 15 s to remember the escape site. After finishing one trial, the mice were kept on dry towels until the next trial. Latency to locate and rest on the hidden platform and the distance traveled (path length) to the hidden platform were recorded to evaluate spatial learning activity. The probe test on the sixth day was performed with the platform removed from the water pool to evaluate long-term spatial memory, where each mouse was released into the water at a position right opposite the escape platform and allowed to swim freely for 60 s. Spatial memory was expressed as the percentage of time spent in the missing platform quadrant and the number of platform location crossings.

NSS test

The NSS test performed before injury was regarded as baseline (n = 7 per group). A 10-point task was used to analyze motor function, balance, alertness, and behaviors at different time points after TBI. One point was given when the task was failed. The total score reflected the severity of brain damage. Tasks used in NSS included the following: (i) ability and initiative to exit a circle of 30 cm in diameter within 3 min; (ii) paresis of upper and/or lower limb of the contralateral side; (iii) alertness, initiative, and motor ability to walk straight; (iv) innate reflex—the mouse will bounce in response to a loud hand clap; (v) physiological behavior as a sign of “interest” in the environment; (vi) ability to balance on a beam of 7 mm in width for at least 10 s; (vii) ability to balance on a round stick of 5 mm in diameter for at least 10 s; (viii) ability to cross a 30-cm-long beam of 3 cm in width; (ix) same task but increased difficulty on a 2-cm-wide beam; and (x) same task but increased difficulty on a 1-cm-wide beam. NSS was evaluated at 1, 3, 7, 14, and 30 days after injury.

Corner test

Sensorimotor function was assessed by the corner test (n = 7 per group). Turning preference was tested before injury as baseline. Two pieces of hardboard (20 cm by 30 cm by 1 cm) were attached together with 30° angle. Each mouse was put between the two-angled hardboard and encouraged to enter the deep corner. Then, both sides of the vibrissae were stimulated simultaneously, and the mouse would turn upward and backward and then face to the open end. Each mouse was given 10 trials, and turning side for each time was recorded. Healthy mice usually turn to either side randomly, while brain unilateral damaged ones tend to turn around in the ipsilateral direction due to injury induced postural asymmetry. The right turn number ratio to a total of 10 turns was shown as a percentage (%).

Rotarod test

Accelerated rotation apparatus (rotarod) is a speed-adjustable rod that is used for assessing motor coordination deficits and balance alterations in rodents. All the mice were pretrained on the spinning rod for 5 days before TBI. The overall training time was 5 min with rod speed accelerating from 4 to 40 rpm. Each mouse was given three trials every day, and those that fell from the rod would be put back to the rotating rod. Most of the mice could stay on the device for 5 min by the last day, and those that failed in this training were excluded. Then, the rotarod test was performed at 1, 7, 30, 45, 60, and 80 days after TBI (n = 7 per group). The latency to fall from the rod (5 min, maximum) was measured and averaged over the three trials.

In vivo oxidative stress

Brains were collected at 1, 7, and 14 days after TBI (n = 5 per group). Mice were anesthetized and perfused with PBS (30 ml). Tissues were homogenized in cold PBS (4 ml) to make the concentration approximately 10% and centrifuged twice at 1000g for 10 min. The lipid peroxide assay was performed according to the instructions of malondialdehyde (MDA) assay kit (S0131, Beyotime). MDA content in tissue homogenate was quantified by standard curve. Tissue homogenate samples were mixed with thiobarbituric acid and antioxidant reagent provided by the kit. After incubating the reaction system in boiling water for 15 min, the solution turned pink with accumulation of precipitate. All samples then cooled down in water, and the MDA content was measured by optical absorption at 532 nm. The accurate calculation of MDA was based on the total amount of protein in each sample, which was tested using enhanced BCA protein assay kit (P0010, Beyotime) following the instructions of the manufacturer.

SOD activity was assessed using the Total Superoxide Dismutase Assay Kit with WST-8 (S0101, Beyotime). According to the protocol, tissue homogenate was mixed with WST-8 and xanthine oxidase in 96-well plates, and then xanthine was added and incubated at 37°C for 20 min. The reaction solution turned yellow, and optical absorption was measured at 450 nm. SOD activity was indicated as the ability of inhibiting the production of superoxide per milligram of protein.

Hydrogen peroxide assay kit (S0038, Beyotime) was used for H2O2 quantification in brain tissue at the same time points previously mentioned. Tissue homogenate was blended with the mixture of Fe2+ and xylenol orange and incubated at room temperature for 20 min. After that, absorbance at 560 nm was tested. The H2O2 level was calculated using standard curves based on the individual protein concentration.

The GSSG level was detected using the GSH and GSSG assay kit (S0053, Beyotime). Brain tissue was homogenized in protein-removing reagent M included in the kit at the ratio of 10 mg/100 μl. Then, the homogenate was centrifuged at the speed of 10,000g at 4°C for 10 min. Supernatant was collected and incubated with GSH-masking reagents at room temperature for 1 hour. GR and 5,5′-dithiobis(2-nitrobenzoic acid) were added in the reaction system and kept at room temperature for 5 min, and then the chromogenic reaction was started by adding NADPH. Twenty minutes later, absorbance was analyzed using a microplate reader at 412 nm. GSSG level was quantified by the standard curve and divided by protein concentration.

•NO is unstable and difficult to test; therefore, it was analyzed by detecting the content of nitrate or nitrite in tissue. The first step was to convert nitrate to nitrite in the presence of nitrate reductase. The second step was adding the fluorescent probe 2,3-diaminonaphthalene to react with nitrite. Meanwhile, NaOH was added to enhance the fluorescent yield, and the fluorescence was measured. Nitrate oxide level was in proportion to the intensity of fluorescence/total protein (in milligrams per milliliter).

Enzyme-linked immunosorbent assay kits for inflammatory cytokine

At days 1, 3, and 14 after the TBI operation (n = 5 per group), mice were cleaned with 10 ml of PBS perfusion, and brain samples were rapidly harvested. Homogenates were centrifuged at 10,000g for 10 min, and supernatant were saved at −80°C for preparation. Supernatant protein concentration was measured and used for inflammation-related cytokine quantification. Enzyme-linked immunosorbent assay (ELISA) kits for IL-1β (ab197742, Abcam) and TNF-α (ab208348, Abcam) were used to detect inflammation levels. These assays were carried out according to the instructions provided by the manufacturer. Each sample was detected twice at least and analyzed using Microsoft Excel 2010 software.

Immunostaining

Brain tissue (n = 3 per group) was made into paraffin sections at 1 and 7 days after TBI. The slides were dewaxed in xylene and gradient ethanol and processed with 0.3% H2O2 and 0.1% Triton X-100 at 37°C for 20 min. After that, all slides were put in antigen-retrieval buffer, kept at 95°C for 10 min, and then allowed to cool to room temperature naturally. Then, the samples were blocked with 10% goat serum at 37°C for 40 min. Anti-GFAP (1:400; ab90601, Abcam), anti–Iba-1 (1:300; ab48004, Abcam), and anti-NeuN (1:800; GTX00837, GeneTex) primary antibodies were added onto each sample and placed at 4°C overnight. The tissue was rinsed in tris-buffered saline with Tween 20 three times and incubated with secondary antibody at room temperature for 1 hour. After counterstaining with 4′,6-diamidino-2-phenylindole (DAPI), the slides were mounted with antifade reagent and photographed with fluorescence microscope (EVOS, AMG).

Real-time quantitative polymerase chain reaction

Total RNA from 100 mg of mouse hippocampus was isolated using RNAiso Plus (catalog no. 9109, Takara). Complementary DNA was made from 2 μg of extracted total RNA with RevertAid Reverse Transcriptase (catalog no. EP0441, Thermo Fisher Scientific). The RT-qPCR was performed with LightCycler 480 SYBR Green I Master (catalog no. 04707516001, Roche Diagnostics) on Applied Biosystems 7500 Real-Time PCR Systems. The reaction was incubated at 95°C for 5 min, followed by 40 cycles of 95°C for 10 s and 60°C for 40 s. Gene-specific primers were as follows: Nrf2, 5′- TCCCATTTGTAGATGACCATGAG-3′ (forward) and 5′- CCATGTCCTGCTCTATGCTG-3′ (reverse); HO-1, 5′- ACAGAGGAACACAAAGACCAG-3′ (forward) and 5′- GTGTCTGGGATGAGCTAGTG-3′ (reverse); NQO-1, 5′-ACGACAACGGTCCTTTCCAGA-3′ (forward) and 5′-CAGAAACGCAGGATGCCACT-3′ (reverse); β-actin, 5′-GACAGGATGCAGAAGGAGATTACT-3′ (forward) and 5′-TGATCCACATCTGCTGGAAGGT-3′ (reverse).

Western blot

Total protein lysate from mouse hippocampus was prepared by radioimmunoprecipitation assay buffer containing protease inhibitors (cOmplete, mini, EDTA-free protease inhibitor cocktail tablets, Roche Diagnostics). The protein concentrations were determined using a BCA protein assay kit (Bio-Rad) and analyzed with SDS–polyacrylamide gel electrophoresis and subsequently transferred to a polyvinylidene difluoride membrane. Anti-Nrf2 (1:1000, #ab62352), anti–HO-1 (1:1000; #10701-1-AP), and anti–β-actin (1:2000; #sc-47778) were purchased from Abcam, ProteinTech, and Santa Cruz Biotechnology, respectively. The horseradish peroxidase (HRP)–conjugated secondary antibody was purchased from Abcam. The signal was detected with Pierce enhanced chemiluminescence Western blotting substrate (catalog no. 32209, Thermo Fisher Scientific) and exposed using UVITEC alliance mini HD6.

IHC staining

The brain sections with antigen retrieval were treated with 3% H2O2 for 15 min and QuickBlock Blocking Buffer (catalog no. P0260, Beyotime, China) for 30 min at room temperature. Then, the sections were incubated with primary antibodies against HO-1 (1:200; #10701-1-AP, ProteinTech) and Nrf2 (1:200; #ab62352, Abcam) overnight at 4°C. After washing twice with PBS, the sections were incubated with biotinylated secondary antibody (1:500; Vector, Burlingame, CA, USA) and HRP-streptavidin reagent (Vector, USA). The 3,3-diaminobenzidine was used as the chromogen. Then, the sections were counterstained with hematoxylin. Images were pictured by a microscope. The ImageJ software was used to quantify positively stained immunoreactive cells.

Biodistribution

O-NZ solution was administered intravenously into healthy or TBI mice at a dose of 100 mg kg−1 at 2 hours after injury. At different time points after injection, the brain and main organs were collected and homogenized to evaluate biodistribution by fluorescence under 455-nm excitation on Maestro EX in vivo imaging system (n = 3). All results of biodistribution were presented by subtracting fluorescence background of the corresponding organs from healthy mice. The uptake of O-NZ in the injury site was observed by brain sections using fluorescence microscope.

Pharmacokinetics and pathology

O-NZ solution was mixed with PBS or FBS and placed at room temperature for 44 hours. Fluorescence was examined using an in vivo imaging system (Maestro EX, Caliper) at the time points of 0, 3.5, 5, 6.5, 20, and 44 hours. O-NZ solution was intravenously injected into C57 mice (n = 3), and blood samples were collected from inner canthus at 5, 15, and 30 min and 1, 6, 12, 24, and 48 hours, respectively. Excretion (n = 3) was collected for 13 hours, and O-NZ in the urine was quantified by fluorescence.

All mice were euthanized by exsanguination at 3 months after injection. Blood was drawn and collected in Eppendorf tubes for hematology analysis on a blood cell counter. For biochemistry analysis, blood samples were centrifuged at 6000 rpm twice to separate the serum and analyzed by biochemistry analyzer (7170S, Hitachi). Main organs including the heart, liver, spleen, lung, and kidney were fixed in 10% formaldehyde for 48 hours. Pathology assessment was performed by hematoxylin and eosin staining.

Statistical methods

Data are presented as means ± SD or SEM. Comparison of means between two groups was accomplished by the Student’s t test. For multiple comparison, one-way analysis of variance (ANOVA) was used to assess difference in means among groups. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus the control group was analyzed by ANOVA. #P < 0.05, ##P < 0.01, ###P < 0.001, and ####P < 0.0001 versus the TBI group was analyzed by ANOVA.