Despite advances in the development of antifungal drugs, there has been an upsurge of cryptococosis infections that poorly respond to fluconazole (first choice drug). Hence, it is paramount to investigate the chemical properties of azole derivatives that are active against resistant C. neoformans. In order to achieve this goal, the susceptibility profile of a clinical isolate of resistant C. neoformans against 33 commercial azole derivatives was evaluated along with their potency (minimum inhibitory concentration, MIC). These data were employed to build SIMCA (soft independent modeling of class analogies) models that pinpoint the importance of electronic features (JGI10) to separate active from inactive compounds and hologram-QSAR models that have good fit but insufficient predictive power (HQSAR, r² = 0.85, q² = 0.35 and r²pred = 0.38). Conversely, 2D QSAR models built from topological descriptors improved the statistical quality (r² = 0.95, q² = 0.86, r²pred = 0.72) and highlight that charge distribution (GGI1) and topological electronegativity (GATS1e and MATS2e) should be modulated to overcome the C. neoformans resistance.
Cryptococosis; azole resistance; 2D QSAR; SIMCA; HQSAR
