Discussion

The BBB permeability of chemicals is of crucial importance for the assessment of neurotoxicity; therefore, we innovatively constructed an assessment method for BBB permeability by computational toxicity and physicochemical property indicators. Using machine learning and QSAR modeling, we found that all six CTPs may alter the permeability of the BBB due to their effects on its structure and function. Among CTPs, HPCTP showed the strongest BBB toxicity. Further analysis confirmed that all six CTPs possess physicochemical characteristics supporting passive diffusion across the BBB. We then applied network toxicology, molecular docking, and molecular dynamics simulations to identify other penetration routes. Previous study has been demonstrated that increased BBB permeability facilitated the transport of exogenous substances into the brain, highlighting the importance of BBB integrity in controlling CNS exposure.41

The demand for high-performance lithium-ion batteries continues to grow, especially in the electric vehicle sector, and this trend has significantly increased the production and use of CTPs.1,2 Recent work revealed that HPCTP disrupts several neurodevelopment-related genes expression such as Mbpa, Syn2a, Gap43, and Gfap, resulting in the impairment of neuronal differentiation, synaptogenesis, and glial maturation, and also downregulates Htr1aa expression, suppressing 5-HT1A receptor function and further inducing depression-like behavior in experimental models.18,20,21 However, these studies do not seem to explain the scientific issue of how HPCTP penetrates the BBB. Previous studies have established a correlation between the likelihood of passive diffusion across the BBB and specific physicochemical properties of molecules.24,42,43 These properties have been utilized in screening strategies for CNS-targeting drugs, such as in evaluating the BBB penetration potential of phenyl-γ-valerolactones (PVLs) and phenylvaleric acids (PVAs).25,44 Thus, our study conducted a BBB toxicity assessment of CTPs via implementing a scoring system based on molecular descriptors derived from the QikProp module in Schrödinger software. This analysis confirms that six CTP analogues can penetrate the BBB via passive diffusion and identifies HPCTP, PFPCTP, HCCTP, and HFCTP as having significant potential for penetrating and damaging the BBB.

To further investigate the molecular mechanism by which HPCTP, PFPCTP, HCCTP, and HFCTP penetrate the BBB, we applied network toxicology methods performing GO and KEGG analyses on their CNS- and BBB-related target genes. Results show that four CTP analogues are linked to the CNS via GO enrichment analysis. More critically, HPCTP exhibits stronger associations than the other three compounds across BP, CC, and MF modules, which is similar to previous study showing HPCTP poses greater potential hazards to living organisms than other CTP analogues.17,18,20 And moreover, HPCTP is clearly related to the ABC efflux transporters and SLC transporters on the BBB, suggesting its potential to modulate key nodes governing BBB permeability. Therefore, our analysis focuses on HPCTP, exploring the active transport of HPCTP by ABC and SLC transporters.

For ABC efflux transporters, target gene prediction revealed that HPCTP associates with genes encoding P-gp/Abcb1, Mrp1/Abcc1, and Bcrp/Abcg2, implying potential disruption of their expression. In addition, predictions from ADMETlab 3.0 suggested that there is a high probability that HPCTP is an inhibitor of P-gp rather than a substrate, and we also discovered that HPCTP has low probability of inhibiting MRP1 and is unlikely to act as a BCRP inhibitor. In general, if a chemical substance meets conditions of downregulating the expression of efflux transporter genes on the BBB, inhibiting their activity and not acting as their substrate, then chemical substance will be less transferred from CNS to bloodstream by efflux transporter system, which may lead to a greater accumulation of chemical substance in the CNS.45 Therefore, we do not exclude the possibility that HPCTP inhibits the P-gp transporter located on the BBB, thereby accumulating more harmful substances in the brain during mixed exposure.

Analysis of HPCTP’s interactions with SLC transporters identified Slc2a1 and Slc6a3 as potential target genes, indicating its capacity to interfere with SLC function and subsequently influence BBB permeability to xenobiotics. Thus, we hypothesize that HPCTP perhaps acts as a substrate for SLC2A1 and SLC6A3 facilitating its own BBB penetration. Subsequently, molecular docking and dynamics simulation results demonstrated strong binding affinity and stability for both HPCTP-SLC2A1 and HPCTP-SLC6A3 complexes, supporting the possibility of HPCTP serving as a substrate for SLC transporters.

Importantly, molecular dynamics simulations further revealed distinct interaction patterns between HPCTP and the two SLC transporters. As shown in Figure 4C, the SLC2A1-HPCTP complex maintained substantially lower RMSD values throughout the 100 ns simulation compared with the SLC6A3-HPCTP complex, indicating higher conformational stability of the SLC2A1-bound system. Similarly, the radius of gyration (Rg) of the SLC2A1-HPCTP complex remained relatively stable during the simulation, whereas the SLC6A3-HPCTP complex exhibited greater structural fluctuations. In addition, the SLC2A1-HPCTP system consistently maintained a higher number of atomic contacts, suggesting tighter and more persistent intermolecular interactions between HPCTP and SLC2A1. Binding free energy analysis further supported these observations. The total binding free energy (ΔGbinding) of the SLC2A1-HPCTP complex was lower than that of the SLC6A3-HPCTP complex, indicating a thermodynamically more favorable interaction. Notably, van der Waals interactions and electrostatic contributions represented the major driving forces stabilizing the SLC2A1-HPCTP complex. Molecular dynamics simulations following virtual point mutations confirmed that the interaction patterns predicted by our molecular docking analysis were reliable. These results collectively suggest that HPCTP may exhibit a stronger interaction tendency toward SLC2A1 than SLC6A3.

Considering that SLC2A1 (GLUT1) is the principal glucose transporter expressed at the BBB and is essential for maintaining cerebral glucose supply, the stronger and more stable interaction between HPCTP and SLC2A1 may have important biological implications beyond BBB penetration. To further validate the biological relevance of the computational predictions, we subsequently performed in vitro experiments using the hCMEC/D3 cells to simulate the BBB. Interestingly, HPCTP exposure induced dose-dependent alterations in glucose uptake, glucose permeability, and intracellular glucose accumulation, accompanied by significant downregulation of Slc2a1 gene expression, which was consistent with the mechanism predicted in our network toxicology analysis. This phenomenon may be associated with the dual effects of HPCTP on glucose uptake and glucose metabolism in hCMEC/D3 cells.

Moreover, the total glucose concentration in the entire system, including the upper chamber medium, lower chamber medium, and cell lysates, also increased in a dose-dependent manner. These findings indicate that HPCTP may interfere with the normal glucose transport function of SLC2A1, thereby disrupting BBB energy metabolism as well as influencing energy supply and metabolic utilization between the BBB and the CNS. Previous studies have demonstrated that BBB dysfunction and alterations in cerebral oxygen and nutrient delivery are closely associated with brain injury progression and neurological impairment.46 Therefore, disruption of SLC2A1-mediated glucose transport by HPCTP may represent an important mechanism contributing to BBB dysfunction and CNS toxicity.

Previous studies have shown that reduced GLUT1 expression at the BBB is associated with BBB breakdown, impaired cerebral glucose uptake, which is linked to neurodegeneration in Alzheimer disease.47 In addition, GLUT1 dysfunction has also been implicated in neurological disorders such as epilepsy due to impaired cerebral energy metabolism.48 Therefore, HPCTP-induced disturbance of SLC2A1 may indicate a potential risk for BBB dysfunction and neurological impairment.

In summary, this study utilized computational toxicology methods to establish a reliable assessment system to evaluate the BBB permeability of CTPs, revealing distinct mechanisms by which CTPs may penetrate the BBB, including passive diffusion and transporter-mediated penetrating. Importantly, our research further demonstrated that HPCTP exposure exerted adverse effects on glucose transport and metabolism, suggesting its potential risk in disrupting CNS energy metabolism. Based on our findings, we proposed that CTP-induced effects on the BBB potentially alter its permeability to other substances. Our research findings underscore the necessity of conducting more in-depth toxicological studies on CTPs, particularly regarding their neurotoxicity. Given the relevance of these compounds to human and environmental health, more in vitro and in vivo studies are essential to fully characterize their risk profiles.

Limitations of the study

The present study has certain limitations, although our in vitro experiments demonstrated that HPCTP affects glucose uptake, glucose distribution, and Slc2a1 expression in hCMEC/D3 cells, our research did not directly demonstrate BBB penetration by HPCTP or its direct impact on BBB integrity. Therefore, further studies employing dedicated BBB permeability assays and in vivo models are required to validate these predictions and clarify the effects of HPCTP on BBB function and permeability.