Comparative In Silico ADME and Drug-Likeness Profiling of Selected Alkaloids and Flavonoids for CNS Drug-Discovery Prioritisation

Background: Central nervous system (CNS) drug discovery is constrained by the difficulty of achieving adequate brain exposure alongside acceptable pharmacokinetic properties. This study compared predicted absorption, distribution, metabolism and excretion (ADME) properties and drug-likeness of selected alkaloids and flavonoids associated with African medicinal plants. Methods: Four alkaloids (vincristine, vinblastine, reserpine and ajmalicine) and four flavonoids (quercetin, kaempferol, luteolin and apigenin) were evaluated computationally. Canonical SMILES and compound identifiers were retrieved from PubChem. SwissADME was used to estimate physicochemical descriptors, drug-likeness filter violations and medicinal-chemistry alerts; pkCSM was used to predict selected ADME endpoints, including blood–brain barrier (BBB)-related parameters. Results: The flavonoids had lower molecular weights (270.24–302.24 g/mol) than the alkaloids (352.43–824.96 g/mol), and all four satisfied the five drug-likeness filters without violations. Ajmalicine was the only evaluated compound predicted to meet both pre-specified pkCSM CNS-related thresholds and all five drug-likeness filters (predicted log BB = 0.510; log PS = −1.989; TPSA = 54.56 Ų). All eight compounds were predicted P-glycoprotein substrates. These are computational predictions and were not experimentally validated. Conclusions: Within this small, selected set, the flavonoids showed more favourable predicted drug-likeness, whereas ajmalicine ranked highest for the reported CNS-related predictions. Differences among compounds are better explained by molecular size, polarity, hydrogen-bonding and predicted transporter interactions than by chemical class. The findings support prioritisation for experimental follow-up, not claims of demonstrated brain penetration, therapeutic efficacy or safety.