Natural products and their derivatives represent a significant source of therapeutic agents, playing a crucial role in the treatment of various human diseases. Notably, an increasing number of active compounds derived from natural plants have been identified, and their respective targets have been systematically investigated [1,2]. Quaternary benzophenanthridine alkaloids (QBAs), a subgroup of benzylisoquinoline alkaloids, have garnered considerable attention owing to their diverse pharmacological activities, which encompass anti-HIV, antimalarial, anti-inflammatory, antibacterial, antituberculosis, and particularly anticancer properties [[3], [4], [5], [6]]. QBAs, such as sanguinarine, chelerythrine, and nitidine, are predominantly found in Sanguinaria canadensis, Chelidonium majus L., and Zanthoxylum nitidum [6] (Fig. 1). Numerous studies have focused on elucidating the anticancer mechanisms of QBAs, primarily through the induction of autophagy or apoptosis [3,7]. These processes are influenced by the downregulation of apoptosis-related genes and proteins (such as NOL3, HRK, BCL2, XIAP, and CLAP2), the alteration of mitochondrial membrane potential, the activation of apoptotic proteins (notably caspase 3) or NF-κB [8,9], the accumulation of reactive oxygen species (ROS) [10], and interactions with nucleic acids [11]. While QBAs exhibit significant anticancer efficacy, evidenced by IC50 values within the micromolar range, advancements in clinical applications have been hindered by suboptimal in vivo activity, limited selectivity, and intricate synthetic routes [6,12,13]. Therefore, it is imperative to rationally design and optimize these structures to investigate the structure-activity relationship (SAR) and achieve simplification of the molecular framework [14]. Several studies have explored the structural simplification of QBAs, including anticancer research focused on simplified derivatives of nitidine [15], and chelerythrine [13]. However, the mechanism of action for the former remains inadequately studied, whereas the latter lacks an N-heterocyclic core, a critical and common feature among such natural products [16,17], which all preclude the development of ideal lead compounds.

Thioredoxin reductase (TrxR, encoded by TXNRD) is an essential component of the thioredoxin system, a crucial redox regulatory system within cellular environments. TrxR receives electrons from NADPH, facilitating its reduction and enabling the transfer of electrons to downstream thioredoxin (Trx, encoded by TXN) [18,19]. Remarkably, TrxR is classified as a selenoprotein, characterized by a highly reactive selenocysteine (Sec) residue located at its C-terminus [20]. The reactivity of Sec greatly exceeds that of cysteine (Cys) under physiological conditions [21,22]. Furthermore, TrxR is frequently overexpressed in tumor cells, rendering it a significant target for the development of anticancer therapeutics [[23], [24], [25], [26], [27], [28]]. Despite its high reactivity, the extensive presence of free thiols, such as glutathione (GSH) and Cys, at millimolar concentrations poses a considerable challenge regarding the selectivity of drugs aimed at targeting TrxR [29,30].

In this study, we first assessed the significance of aromatic scaffolds and quaternary nitrogen in the core of N-heterocycles by optimizing the structure of QBAs. We subsequently refined the structure from the four-membered ring QBAs to the three-membered ring phenanthridine. Utilizing tertiary nitrogen phenanthridine as a control molecule, we synthesized a series of compounds and evaluated their cytotoxicity against HeLa, HepG2, and A549 cells, as well as their inhibitory effects on TrxR in vitro. Notably, compound 6f exhibited the highest cytotoxicity against A549 cells and demonstrated significant inhibition of TrxR in vitro, while showing no inhibitory activity against the U498C TrxR mutant (Sec→Cys) and glutathione reductase (GR). Given the prevalence of free thiols in vivo, which may interfere with the inhibition of TrxR by 6f, we conducted experiments where 6f was incubated with a 20-fold excess of GSH or Cys under identical conditions, producing unexpectedly negative results. Further investigations revealed that TrxR can be selectively inhibited by 6f, which additionally induces apoptosis through the generation of intracellular oxidative stress. Moreover, pronounced tumor regression was observed in nude mice with non-small cell lung cancer (NSCLC) following treatment with 6f. Therefore, we have optimized the structure of QBAs and identified 6f as a candidate with remarkable inhibitory effects on NSCLC. Importantly, 6f demonstrates significant selectivity for TrxR, without reacting with GSH and Cys. Our findings not only present a highly selective scaffold for TrxR inhibitors but also contribute to a candidate for the development of effective benzophenanthridine-based anticancer agents.