Flow chart of experimental procedures and characterization of fullerenol
The overall experimental study design was illustrated in Fig. 1A, including information about the timeline of the procedures. Water-soluble fullerenol was difficult to be characterized because of its amorphous structure. Fullerenol exhibited a circular or rectangular morphology with particle size of about 100 nm as revealed by TEM (Fig. 1B). It was found that the size distribution of single phase fullerenol obeyed normal distribution, and the surface zeta potential was -24.98 ± 3.85 mV (Fig. 1C).
Identification of fulleronol on animal models
The biological distribution of fullerenol in a rat (i.p., 5 mg/kg) was detected by MALDI-TOF–MS. The mass peak of water-soluble fullerenol [C60(OH)24] after acidification and purification was at m/z 1200 [M + 4H2O] (Fig. 2A), while in the main tissues (brain, liver, kidney and spleen) the most abundant ion might be centered at m/z 1055 [M-4H2O-H]− and corresponded to a singly charged ion generated through fullerenol (Fig. 2B-E), indicating that fullerenol might widespread distribute and accumulate in important organs within the body by passing the blood–brain barrier[21]. Compared with rats without lead treatment, the lead content in blood (Total df = 35, F = 19.43, p < 0.001) and hippocampus (Total df = 36, F = 15.72, p < 0.001) of lead-exposed rats had increased significantly, proving that an animal model of chronic lead poisoning had been successfully made (Fig. 2F).
Fullerenol improved hippocampus-dependent spatial learning and memory
In open field test, the number of crossing squares (Total df = 51, F = 2.32, p > 0.05; Total df = 51, F = 0.54, p > 0.05) and rearing (Total df = 42, F = 3.92, p > 0.05) were not significantly changed by fullerenol and lead (Fig. 3A-B), which strongly indicating that the exercise capacity and exploratory activity of rats after various treatments were normal.
In Morris water maze test, lead-exposed animals presented a longer latency to find the platform than non-lead-exposed rats during the training period (Corrected Total df = 222, Model F = 15.11, p < 0.001, Fig. 3C). During the testing period, compared with control rats, lead-exposed rats spent less time in the correct quadrant (Total df = 33, F = 33.67, p < 0.001, Fig. 3D) and took more time to find the correct area (Total df = 41, F = 11.14, p < 0.01, Fig. 3E). However, it was improved in the fullerenol-intervene group (p < 0.01 and p < 0.001 compared with the lead-exposed group, Fig. 3D, E). Meanwhile, there was no significant difference in the average swimming speed among each group (Total df = 51, F = 0.22, p > 0.05, Fig. 3F). These results showed that fullerenol could improve hippocampus-dependent cognition and protect rats against lead-exposure induced impairment in spatial learning and memory.
Fullerenol enhanced long-term synaptic plasticity in the hippocampus
The schematic diagram is in the upper left corner of Fig. 4A. fEPSP slope is calculated as the slope of the first positive wave peak. And PS amplitude is calculated as the distance between the wave bottom and the midpoint of the crest line of the two positive peaks. As shown in Fig. 4A-B, the I/O curve corresponding to fEPSP slope was enhanced in the fullerenol-exposed group (Corrected Total df = 314, Model F = 26.61, p < 0.05), while PS amplitude had a slightly increasing trend in both of the two fullerenol treatment group (Corrected Total df = 294, Model F = 21.89, p > 0.05). In follow-up experiments, the stimulation intensity was adjusted to evoke potentials that 50% of the maximal PS amplitude, expressed as 0.4 mA.
As shown in Fig. 4C, D, the PPF ratio determined using the measurement of fEPSP2/fEPSP1 and PS2/PS1 had no obvious changes among groups. In addition, as shown in Fig. 4E, fEPSP slope after HFS in the fullerenol-exposed group (126.59 ± 0.94%) was higher than that in the control group (114.03 ± 0.88%, Corrected Total df = 511, Model F = 27.31, p < 0.001). Meanwhile, in Fig. 4F, PS amplitude was increased in the fullerenol-exposed group (266.95 ± 1.73%, Corrected Total df = 591, Model F = 32.49, p < 0.001) and decreased in the lead-exposed group (182.4 ± 2.05%, p < 0.01) by comparing with the control group (215.1 ± 1.73%). Importantly, the suppression in the lead-exposed group vanished in the fullerenol-intervene group (fEPSP slope: 122.38 ± 0.7%, p < 0.05; PS amplitude: 229.75 ± 1.36%, p < 0.001; compared with the lead-exposed group). These results indicated that fullerenol could enhance synaptic plasticity in the DG area and alleviate the damage caused by lead toxicity to some degree.
EPSP slope is calculated as the slope of the positive wave peak, the schematic diagram is in the upper left corner of Fig. 5A. Additionally, as shown in Fig. 5A, the basal synaptic transmission in the Sch-CA1 pathway was decreased by lead (Corrected Total df = 467, Model F = 35.65, p < 0.01) and increased by fullerenol (p < 0.001), as manifested by in the shift of the I/O function curve. In PPF and LTP experiments, the intensity of single pulses evoking 50% of the maximal PS amplitude was expressed as 0.6 mA.
As shown in Fig. 5B, compared with the control group (150.45 ± 4.88%), the peak of the PPF ratio curve was significantly reduced in the lead-exposed group (122.48 ± 1.13%, Total df = 42, F = 9.43, p < 0.001) and rescued in the fullerenol-intervene group (145.83 ± 4.25%, p < 0.01, compared with the lead-exposed group), indicating that the lead-induced short-term depression could be altered by fullerenol. Furthermore, as shown in Fig. 5C, the fEPSP slope after HFS was abated by lead (125.32 ± 0.91%, Corrected Total df = 2271, Model F = 41.35, p < 0.001, compared with the control group 148.4 ± 1.65%) and raised by fullerenol (189.63 ± 0.52%, p < 0.001), while the reduction due to lead disappeared in the fullerenol-intervene group (149.83 ± 0.56%, p < 0.001). Notably, this maintenance of synaptic plasticity was a long-lasting strengthening, and lasted at least for 4 h. These results indicate that long-term enhancement of synaptic efficacy occurred in the CA1 area, and fullerenol could change lead-impaired synaptic transmission in the hippocampus.
Fullerenol altered the PSD-dependent structure in hippocampal CA1 primary neurons
In Fig. 6A, using Golgi staining, dendritic spines were clearly marked and the number of the second branches of dendrites was accurately counted. The mixed effect model was used, while the different slices in the same animal were set as the random effect. The significant value of the random effect was 0.945, and it eliminated the type I error. In Fig. 6B, compared with the control group (8.64 ± 0.11/10 µm), the number was significantly decreased in lead-exposed group (7.81 ± 0.08/10 µm, Total df = 626, Model F = 147.02, p < 0.001), and this trend had been different in the fullerenol-intervene group (8.52 ± 0.06/10 µm, p < 0.001, compared with the lead-exposed group). The dendritic spines of neurons are closely related to the formation of PSD plaques. In the center of Fig. 6C, PSD were clearly visible under the electron microscope, and the number of PSD in the lead-exposed group (12.2 ± 0.49/1µm3) was significantly less than that in the control group (15 ± 0.77/1µm3, Total df = 67, Model F = 4.55, p < 0.05), but this decrease was significantly attenuated in the fullerenol-intervene group (14.84 ± 0.73, p < 0.01, compared with the lead-exposed group, Fig. 6D). Meanwhile, in Fig. 6E, the level of PSD95 protein was similar to the above quantitative change (Total df = 11, Model F = 10.88). Calcium/calmodulin-dependent protein kinase II (CaMKII) is implicated in LTP, and some studies have shown that there is an increase in CaMKII activity directly in the PSD of dendrites after LTP induction. In Fig. 6F, exposure to lead significantly reduced the pCaMKIIα/CaMKIIα ratio (Total df = 11, Model F = 10.59, p < 0.05), but the reduction induced by lead disappeared in the fullerenol-intervene group (p < 0.05, compared with the lead-exposed group). These findings indicated that the treatment with fullerenol in our study did up-regulate the number of dendritic spines, the level of PSD95 and the activity of CaMKIIα, which might enhance synaptic efficacy.
This protective effect of fullerenol was independent on the reduction–oxidation pathway
According to the previous study using cultured cells, the protective effect of fullerenol was associated with the redox level [20]. In order to assess the in vivo protective effect of fullerenol, we determined the redox state in hippocampal tissues. As shown in Fig. 7, we found that the level of H2O2 increased from 3.529 ± 0.257 μmol/g in the control group to 4.86 ± 0.387 μmol/g in the lead-exposed group (Total df = 49, F = 8.59, p < 0.05), while the total SOD activity, the total antioxidant capacity and the total GSH concentration decreased from 2.63 ± 0.47 units, 0.94 ± 0.03 mmol/g and 24.09 ± 1.74 μM in the control group to 1.37 ± 0.22 units, 0.81 ± 0.01 mmol/g and 17.79 ± 0.65 μM in the lead-exposed group (Total df = 39, F = 8.76, p < 0.05, Total df = 39, F = 15.47, p < 0.001; Total df = 29, F = 7.41, p < 0.01). However, there were no significant changes between the lead-exposed group and the fullerenol-intervene group (p > 0.05). These results indicated that potentiation of spatial learning and memory by fullerenol might be not dependent on the reduction–oxidation pathway.