Visual inspection (Figure 1A) shows a pale yellow solution displaying the Tyndall phenomenon, characteristic of colloidal systems. Morphological analysis via TEM (Figures 1B,C) confirms that the CrCi-CDs are well-dispersed nanospheres with uniform dimensions. High-resolution TEM (HRTEM) analysis (Figure 1D) provided insights into the atomic arrangement of the nanodots, revealing well-defined lattice fringes. The measured interlayer d-spacing of 0.228 nm is in excellent agreement with the characteristic lattice parameters of graphitic carbon, confirming the crystalline nature of the CrCi-CDs.

As shown in Figure 1E, the amorphous nature of the CrCi-CDs was evidenced by an XRD diffraction peak centered at 2θ = 29.399°. Optical characterization (Figure 1F) revealed a smooth UV-Vis absorption curve with a weak shoulder around 320 nm, attributed to n-π* (C=O) or π-π* (C=C) transitions. Fluorescence analysis indicated that the CrCi-CDs emit bright blue light under a 365 nm UV lamp, with a peak emission at 402 nm when excited at 334 nm (Figure 1G).

Surface functionalization was corroborated by FTIR (Figure 1H) and XPS analyses. The FTIR spectrum displayed characteristic peaks for O-H/N-H (3,440 cm⁻¹), C-H (2,924 cm⁻¹), and C=O (1,630 cm⁻¹) groups. Further peaks at 1,401 cm⁻¹ (C-N/N-H/C-H), 1,262 cm⁻¹ (C-OH), 1,101 cm⁻¹ (C-O-C), and 863 cm⁻¹ (aromatic C-H) suggest a surface rich in hydroxyl, amino, and carboxylate moieties.

Elemental composition was quantified via XPS survey scans (Figure 1I), yielding 70.7% C, 6.13% N, and 23.17% O at binding energies of 285.08, 400.08, and 532.08 eV, respectively. High-resolution spectra (Figures 1J–L) further resolved the bonding states: C1s deconvolution showed C-C, C-N, C-O, and C=O species; O1s fitting revealed C-O (531.0 eV) and C=O (532.0 eV) bonds; and N1s analysis confirmed the presence of C-N (399.5 eV) and C=N (400.5 eV) structures. These results align with prior data, confirming the successful and stable preparation of CrCi-CDs.

As illustrated in the experimental timeline (Figure 2A), mice received oral CrCi-CDs for 7 days prior to epilepsy induction. Figure 2B demonstrates that CrCi-CDs intervention reduced epilepsy severity across all three epilepsy models. Notably, the high-dose CrCi-CDs and positive control groups (VPA) exhibited the most significant alleviation of epilepsy symptoms.

To evaluate spatial learning and memory, the PTZ model was subjected to the Morris water maze. Representative swimming trajectories (Figure 2C) revealed disorganized search patterns in the PTZ model group, which were regularized by drug intervention. Quantitatively, the PTZ model group exhibited significantly prolonged escape latency and increased swimming distances compared to controls (P < 0.01, Figures 2D,E). However, intervention with VPA and high-dose CrCi-CDs significantly mitigated these learning deficits from day 1 to day 4 (P < 0.01).

In the spatial exploration test on day 5 (Figure 2F), the PTZ model mice showed poor memory retention, predominantly searching quadrants opposite the target. Statistical analysis confirmed that the PTZ model group had significantly increased latency and fewer platform crossings compared to controls (P < 0.01, Figures 2G,H). Conversely, VPA and high-dose CrCi-CDs effectively reversed these impairments, significantly reducing latency (P < 0.01) and increasing platform crossings (P < 0.05 and P < 0.01, respectively), demonstrating the restoration of spatial memory.

Histological analysis of the PTZ-induced model (Figures 3A,B) revealed severe structural disarray in hippocampal tissue compared to controls. The PTZ model group was characterized by widened intercellular spaces, cellular swelling, cytoplasmic vacuolization, and evident neuronal necrosis. In contrast, these morphological defects were markedly alleviated following the administration of VPA and various concentrations of CrCi-CDs. Furthermore, Nissl staining (Figures 3C,D) demonstrated that PTZ administration caused a marked depletion of cytoplasmic Nissl bodies, a defect that was effectively reversed by CrCi-CDs and VPA treatment. Consistent neuroprotective effects were observed in the PILO and PNC epilepsy models (Figures 4, 5), confirming that CrCi-CDs possess efficacy in mitigating neuronal damage across different epilepsy paradigms.

The massive release of cytokines by glial cells and neurons triggers inflammatory cascades that are known to lower the seizure threshold and aggravate neurotoxicity (Rana and Musto, 2018). In this study, we specifically focused on the expression profiles of TNF-α, IL-1β, IL-6, and IL-18 (Figures 6A–D). Our results revealed that regardless of the chemical convulsant used (PTZ, PILO, or PNC), the induction of epilepsy was invariably accompanied by a sharp upregulation of these pro-inflammatory markers (P < 0.01). This consistent spike across different models suggests a common inflammatory pathway underlying chemically induced seizures. Crucially, CrCi-CDs treatment interrupted this pathological cycle. Following administration, we observed a progressive, dose-dependent downregulation of all four cytokines. The high-dose CrCi-CDs group, in particular, exhibited a potent anti-inflammatory capacity, effectively suppressing the overexpression of TNF-α, IL-1β, IL-6, and IL-18. This data suggests that CrCi-CDs do not merely mitigate symptoms but actively improve the microenvironment of neural tissue by resolving acute neuroinflammation.

Excessive reactive oxygen species (ROS) exacerbate epileptic brain injury by depleting antioxidant systems (e.g., SOD) and generating toxic intermediates like MDA, which serves as a reliable marker for oxidative damage (Sun et al., 2022; Wang et al., 2021). As shown in Figures 7A–C, across the three epilepsy models, the model groups exhibited significantly reduced SOD activity and elevated MDA levels compared to controls (P < 0.01). However, intervention with VPA and CrCi-CDs effectively normalized these markers. Specifically, VPA and high-dose CrCi-CDs significantly reversed the oxidative stress indices (P < 0.01), with the high-dose group demonstrating the most potent antioxidative efficacy among the CrCi-CDs treatments.

The disruption of the excitatory/inhibitory balance, characterized by dysregulated Glu and GABA expression, is a central mechanism in epilepsy. As illustrated in Figures 7D,E, the model groups in all three induction protocols showed significantly upregulated Glu and downregulated GABA levels compared to controls (P < 0.01). CrCi-CDs intervention dose-dependently corrected this imbalance. Notably, both VPA and high-dose CrCi-CDs significantly decreased Glu and increased GABA levels (P < 0.01). These biochemical findings align with the behavioral improvements, confirming that CrCi-CDs effectively ameliorate the Glu/GABA imbalance and mitigate neuronal excitotoxicity.

Seizures induce excitotoxicity and inflammation, leading to neuronal damage (Sarlo and Holton, 2021; Vezzani et al., 2015). Validated by histological findings (Figures 3–5), we further explored the underlying mechanisms in the PTZ model by analyzing protein expression in the EAAT2, NF-κB, and apoptotic pathways. Regarding glutamate transport (Figure 8A), the model group exhibited significantly elevated EAAT2 expression compared to controls (P < 0.01). This likely reflects a compensatory upregulation in response to excessive synaptic glutamate. CrCi-CDs intervention downregulated EAAT2 expression, with the high-dose group showing a significant reduction (P < 0.05), indicating that CrCi-CDs effectively mitigated synaptic glutamate accumulation and neuronal excitability. We further assessed the NF-κB inflammatory pathway (Figure 8B). The model group showed significantly increased levels of p-IKBα and p-p65 (P < 0.01), indicating pathway activation. High-dose CrCi-CDs markedly decreased these phosphorylation levels (P < 0.05), thereby suppressing NF-κB signaling and the downstream inflammatory cascade. Finally, we evaluated neuronal apoptosis (Figure 8C). The model group displayed significantly elevated pro-apoptotic markers (Bax, Cleaved-caspase-3) and reduced anti-apoptotic Bcl-2 (P < 0.01). CrCi-CDs treatment significantly reversed these trends (P < 0.05), upregulating Bcl-2 while downregulating Bax and Caspase-3. These results confirm that CrCi-CDs exert neuroprotective effects by inhibiting epilepsy-induced neuronal apoptosis.

The study utilized PTZ, VPA and sodium pentobarbital from Sigma-Aldrich (United States) and PILO from APExBIO Technology (United States). Domestically, PNC and 1,000 Da dialysis membranes were attained from North China Pharmaceutical Co., Ltd. and Beijing Ruida Henghui Technology Development Co., Ltd., respectively. All experimental solutions were prepared using deionized water (DW) to ensure consistency across the study.

CrCi-CDs were fabricated from human hair in accordance with our previously described methodology (Zhang et al., 2021). Briefly, cleaned hair was desiccated at 60 °C for 24 h prior to being placed in crucibles. The samples underwent a specific thermal treatment: a 20-min temperature ramp to 70 °C, followed by carbonization at 350 °C for 1 h. Upon cooling to room temperature, the carbonized residue was ground into a fine powder and suspended in deionized water (DW). The suspension was subsequently boiled at 100 °C for 1 h to extract the CDs. To purify the product, the mixture was filtered through a 0.22 µm membrane, and the filtrate was dialyzed against DW for 72 h (MWCO: 1,000 Da) to obtain the final CrCi-CDs solution. Figure 9 briefly illustrates the preparation process of CrCi-CDs.

Spectrofluorimetric data were acquired with a Shimadzu RF5-5301 PC instrument, and UV-visible profiles were obtained through a TU-1991 spectrophotometer. Functional group composition was monitored via FTIR (Nicolet 6700) across a spectral range of 500–4,000 cm⁻¹. To evaluate the physical dimensions and structural features, TEM images were captured using a JEM-2100F microscope (operating at 200 kV). These analyses provided comprehensive insights into the optical and structural attributes of the resultant carbon dots.

Male C57BL/6 mice (20.0 ± 2.0 g) were acclimatized under standard laboratory conditions: a controlled temperature of 24.0 °C ± 1.0 °C, 55%–65% relative humidity, and a 12-h light/dark cycle. Standard chow and water were provided ad libitum.

The animals were stratified into three independent batches to establish different epilepsy models (Three batches, 42 mice in each batch). Within each batch, mice were randomized into six groups (n = 7): Control, Model, positive control (VPA, 200 mg/kg), and three CrCi-CDs treatment groups (low dose group: CrCi-CDs 1.5 mg/kg; medium dose group: CrCi-CDs 3 mg/kg; high dose group: CrCi-CDs 6 mg/kg). The dosing regimen for CrCi-CDs was selected based on our previous study demonstrating neuroprotection in ischemic stroke (Zhang et al., 2021), combined with the conversion of the standard clinical dosage of Crinis Carbonisatus in Traditional Chinese Medicine to murine equivalents. All treatments were administered via intragastric gavage once daily for seven consecutive days, while the Control and Model groups received an equivalent volume of saline.

On Day 7, 60 min after the final prophylactic treatment, epilepsy were induced via intraperitoneal (i.p.) injection. The three cohorts were challenged with PTZ (65 mg/kg), PILO (280 mg/kg), or PNC (4.2 g/kg), respectively (Janisset et al., 2022; Curia et al., 2008; Akdogan et al., 2008). All control groups received i.p. saline injections. Animals in each group were monitored continuously for 30 min post-injection to observe the severity of epilepsy and quantified according to the Racine scale (Racine, 1972).

The PTZ-induced epilepsy model, a classic paradigm in epilepsy research, was selected for this assay due to its high value in primary drug screening and mechanistic investigation. Since PTZ-induced seizures typically result in hippocampal excitotoxicity and subsequent cognitive impairment, the Morris water maze was employed to evaluate the preventive neuroprotective efficacy of CrCi-CDs against these seizure-associated cognitive deficits.

A circular pool with dimensions of 90 cm (diameter) by 50 cm (height) was employed for the water maze tests. The water temperature was stabilized at 22 °C–24 °C, and all experimental data were recorded using an EthoVision XT automated monitoring system. The pool was divided into four quadrants, with visual cues mounted on the walls to facilitate spatial orientation. A hidden escape platform (9 cm diameter) was submerged 1 cm below the water surface in the target quadrant. lighting was kept soft and diffuse to prevent reflections.

Following the behavioral test, after a 24-h recovery period, mice in each group were anesthetized via intraperitoneal injection of sodium pentobarbital (40 mg/kg). Blood samples were then rapidly collected through the retroorbital plexus. Mice were then humanely euthanized by decapitation under anesthesia. Whole brains were rapidly harvested on ice, rinsed with pre-chilled saline to remove residual blood, blotted dry, and weighed. The brain tissue was then bisected along the sagittal midline. The left hemisphere was snap-frozen in liquid nitrogen and stored at −80 °C for biochemical analysis, while the right hemisphere was fixed in 4% paraformaldehyde for histopathological examination.

Fixed brain tissues were dehydrated through a graded ethanol series (50%–100%), cleared in xylene, and embedded in paraffin wax. Coronal sections (3 μm thick) were cut using a microtome. Prior to staining, all sections underwent standard deparaffinization and rehydration through a series of xylene and graded alcohols.

Hematoxylin and Eosin (H&E) Staining: Rehydrated sections were stained with hematoxylin for 5 min, differentiated in warm water for 10 min, and counterstained with eosin for 2 min. After dehydration and mounting with neutral resin, general neuronal morphology was examined under a light microscope.

Nissl Staining: Sections were incubated in toluidine blue solution at 37 °C for 4 h. Following rinsing and dehydration, the slides were cleared in xylene and mounted. The integrity of Nissl bodies in the hippocampus was evaluated microscopically to assess neuronal damage.

After mechanically dissociating the frozen tissue in a pre-chilled environment, the brain homogenate was centrifuged to obtain the supernatant. The concentrations of amino acid neurotransmitters (GABA, Glu) and inflammatory cytokines (TNF-α, IL-1β, IL-6, IL-18) were quantified using specific ELISA kits purchased from Beijing Dongge Boye Biotechnology Co., Ltd. (Beijing, China). All procedures were conducted strictly according to the manufacturer’s instructions, and optical density was measured at 450 nm. Simultaneously, oxidative stress markers were assessed using biochemical assay kits from the same manufacturer. Specifically, Superoxide Dismutase (SOD) activity was determined via the WST-8 method (Ukeda et al., 1999), while Malondialdehyde (MDA) levels were measured using the Thiobarbituric Acid (TBA) method (Ohkawa et al., 1979). Protein concentrations were determined using the Bradford assay to normalize the data.

Brain tissues were lysed in RIPA buffer containing protease inhibitors, and the resulting protein concentration was measured using a BCA kit (Bradford, 1976). Equal protein quantities were fractionated on SDS-PAGE gels and transferred to polyvinylidene fluoride (PVDF) membranes. After 1.5 h of blocking with 5% skim milk powder, the membrane was incubated overnight with the primary antibody on a rocking incubator, followed by incubation with the secondary antibody at room temperature for 1 h. Following TBST rinses, target protein signals were detected via enhanced chemiluminescence and quantified using a digital imaging analysis system.

Data are expressed as mean ± standard deviation (SD). Statistical comparisons were performed using GraphPad Prism software (Version 9.0). For comparisons among multiple groups, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s multiple comparisons test. For non-normally distributed data (e.g., behavioral seizure scores), the Kruskal–Wallis test followed by Dunn’s multiple comparisons test was applied. A P < 0.05 was considered statistically significant. The criteria for significance were defined as *P < 0.05, **P < 0.01, and ***P < 0.001 vs. the corresponding control or model group as indicated in each figure and table. For in vivo animal experiments, each group consisted of seven biological replicates (n = 7).