Exploring bat-inspired cyclic tryptophan diketopiperazines as ABCB1 Inhibitors

The study addresses the pressing problem of chemotherapy resistance driven in part by overexpression of the ABCB1 (P-glycoprotein/MDR1) efflux transporter. Despite decades of efforts, first-, second-, and third-generation ABCB1 inhibitors have failed clinically due to toxicity, off-target effects, and pharmacokinetic issues. The authors propose a biologically inspired strategy: leverage bats, which exhibit low cancer incidence and high, ubiquitous ABCB1 expression, to identify endogenous metabolites that might serve as safe, competitive ABCB1 inhibitors. They hypothesize that tryptophan-derived structures preferred by ABCB1 in bats could be adapted as competitive inhibitors in human ABCB1, thereby overcoming drug resistance by blocking efflux of chemotherapeutics.

- Clinical failures of ABCB1 inhibitors: First-generation (verapamil, cyclosporine A) required high doses and caused toxicity; second-generation (dexverapamil, PSC 833) improved affinity but had toxicity and unpredictable PK; third-generation (tariquidar, VX-710, GF120918) had higher potency but off-target effects and other issues in trials. - Structural advances: Cryo-EM and DEER studies revealed ABCB1 conformations and distinct binding sites. Competitive inhibitors often occupy both the central cavity and the access tunnel, interacting with hydrophobic residues (e.g., phenylalanines F239, F303, F994), providing a rationale for structure-guided inhibitor design. - Evolutionary context: ABCB1 shows positive selection in bats; high expression likely supports detoxification and efflux of metabolic by-products from flight metabolism, motivating exploration of natural metabolite scaffolds (e.g., tryptophan derivatives) for safer inhibitors. - Prior natural inhibitor observations: Foods like garlic, green tea, ginseng, and fruits can inhibit ABCB1 activity, but active components and therapeutic utility remain unclear.

- Cell models: Bat PaKiT03 (Pteropus alecto kidney) with CRISPR ABCB1 knockout clones; human HEK293T with stable ABCB1 overexpression and control; HCT-15 colorectal cancer (endogenous ABCB1); BeWo (ABCG2); HEK293T and HCT-15 with ABCC1 CRISPR KO; HCT-15 ABCB1 and ABCC1 double KO. - Metabolomics: CE-MS profiling comparing parental vs ABCB1 KO PaKiT03 cells; pathway analysis (Ingenuity Pathway Analysis) to identify altered pathways and intracellular tryptophan levels. - Biochemical assay: ATPase activity of purified, reconstituted human ABCB1 in presence of tryptophan, Cyclo-(L-Trp-L-Trp), and verapamil, measuring ADP via HPLC. - Efflux/accumulation assays: Flow cytometry quantifying intracellular accumulation of ABC transporter substrates: rhodamine 123 and doxorubicin (ABCB1), CMFDA (ABCC1), and Hoechst 33342 (ABCG2). Pre-treat cells with candidate compounds (tryptophan; cyclic dipeptides: Cyclo-(L-Trp-L-Trp), Cyclo-(L-Leu-L-Trp), Cyclo-(L-Trp-L-Pro), Cyclo-(L-Leu-L-Pro); methylated and benzylated derivatives; verapamil; tariquidar; Ko143). - DNA damage readout: Western blot for γH2AX after doxorubicin with or without inhibitors to assess functional ABCB1 inhibition via increased intracellular doxorubicin. - Compound design and synthesis: Structure-guided modifications of Cyclo-(L-Trp-L-Trp). Synthesis of Cyclo-(L-1-methyl-Trp-L-1-methyl-Trp) and regioselective benzylation to generate C3N-Dbn-Trp2 (C3-N1-dibenzyl-exo-pyrroloindoline-cyclic-L-tryptophan-L-tryptophan); purification and characterization (details in Supplementary). - Computational modelling: Autodock Vina docking of pairs of ligands into human ABCB1 (PDB 6QEX); membrane-embedded triplicate 100 ns MD simulations (CHARMM36m/CHARMM-GUI) with loop modelling vs truncated loop; analysis of protein–ligand and ligand–ligand interactions, and MM-PBSA binding free energies; LigPlot+ interaction diagrams. - Cell viability and proliferation: Viakrome 405 staining (48 h, ±1 µM doxorubicin) and ATPlite luminescence (24 h, ±2.5 µM doxorubicin); xCELLigence real-time impedance over 60–85 h for proliferation under inhibitor-only conditions. - Gene editing: CRISPR/Cas9 KO strategies and validation (sequencing, Western blot; functional sorting by substrate accumulation). - Statistics: Unpaired two-tailed Student’s t-test; triplicates or indicated replicates for assays.

- Metabolomics: ABCB1 KO PaKiT03 cells showed downregulation of tryptophan degradation pathways and elevated intracellular L-tryptophan, implicating tryptophan-related metabolites as ABCB1-relevant. - ATPase assay: L-tryptophan (25 mM) did not stimulate ABCB1 ATPase; Cyclo-(L-Trp-L-Trp) (2 mM) did stimulate ATPase, indicating recognition by ABCB1; verapamil remained a stronger stimulator. - Efflux inhibition in bat cells: Cyclo-(L-Trp-L-Trp) at 1 mM increased rhodamine 123 and doxorubicin accumulation; cyclic dipeptides lacking two tryptophans showed minimal/no effect. γH2AX increased upon doxorubicin when co-treated with verapamil or Cyclo-(L-Trp-L-Trp); ABCB1 KO cells had high doxorubicin accumulation and γH2AX regardless of treatment. - Human ABCB1 inhibition: In HEK293T ABCB1-expressing cells and HCT-15 cells, Cyclo-(L-Trp-L-Trp) increased doxorubicin accumulation; related dipeptides lacking two tryptophans did not. γH2AX increased with Cyclo-(L-Trp-L-Trp) similar to verapamil. - Computational insights: Two molecules of Cyclo-(L-Trp-L-Trp) or Cyclo-(L-1-methyl-Trp-L-1-methyl-Trp) can bind concurrently—one in central cavity and one in access tunnel near hydrophobic residues (F239, F303, F994). Methylation improved fit and interactions; MM-PBSA indicated stronger binding: Cyclo-(L-Leu-L-Pro) −33.7 ±2.0 kcal/mol; Cyclo-(L-Trp-L-Trp) −56.2 ±1.7; Cyclo-(L-1-methyl-Trp-L-1-methyl-Trp) −46.9 ±3.1; C3N-Dbn-Trp2 −62.7 ±0.7. C3N-Dbn-Trp2 also showed increased ligand–ligand hydrophobic contacts (6.8 ±3.2). - Potency improvement by methylation: Cyclo-(L-1-methyl-Trp-L-1-methyl-Trp) inhibited ABCB1 at lower concentrations (as low as 0.03–0.25 mM) vs Cyclo-(L-Trp-L-Trp) (∼1 mM) in PaKiT03, HEK293T ABCB1, and HCT-15 cells; enhanced ATPase stimulation vs non-methylated. - Benzylation (C3N-Dbn-Trp2): At 10 µM, C3N-Dbn-Trp2 inhibited rhodamine 123 efflux comparably to 10 µM verapamil in PaKiT03; far more effective than methylated analogue (≥125 µM). Minor benzylated isomers were less active. - IC50 comparisons (Rh123 efflux inhibition): HEK293T ABCB1: C3N-Dbn-Trp2 5.9 µM; verapamil 20.5 µM; tariquidar 0.016 µM. HCT-15: C3N-Dbn-Trp2 2.2 µM; verapamil 7.2 µM; tariquidar 0.058 µM. PaKiT03: C3N-Dbn-Trp2 0.9 µM; verapamil 1.0 µM. - Chemo-sensitization: Co-treatment with C3N-Dbn-Trp2 or verapamil reduced viability further in ABCB1-expressing HEK293T and HCT-15 cells treated with doxorubicin (to ~50%), mimicking control/ABCB1 KO cells treated with doxorubicin alone; C3N-Dbn-Trp2 alone did not impair proliferation (xCELLigence) or viability. - Selectivity: At 10 µM, C3N-Dbn-Trp2 did not inhibit ABCC1 (CMFDA efflux) in HEK293T or HCT-15 ABCB1 KO cells; partially inhibited ABCG2 (Hoechst 33342 efflux) in BeWo cells versus strong inhibition by Ko143. Modeling supports most stable interaction with ABCB1.

The work validates a bat-inspired, tryptophan-based scaffold to design competitive ABCB1 inhibitors. Starting from the observation that ABCB1 KO bat cells accumulate tryptophan and downregulate its degradation pathway, the authors reasoned that tryptophan-derived structures might fit ABCB1 binding regions. Functional assays confirmed that the cyclic tryptophan diketopiperazine specifically inhibits ABCB1-mediated efflux, increasing intracellular chemotherapeutic accumulation and DNA damage. Structure-guided design informed by cryo-EM-derived binding concepts (central cavity plus access tunnel occupancy) and MD simulations led to modifications (N1-methylation and C3 benzylation) that enhanced hydrophobic contacts and binding stability. The benzylated derivative C3N-Dbn-Trp2 achieved sub- to low-micromolar IC50s and outperformed verapamil in human cells, with no detectable cytotoxicity when used alone. Partial activity against ABCG2 and lack of ABCC1 inhibition indicate promising selectivity, though further optimization is needed. These findings address the problem of ABCB1-mediated drug resistance by demonstrating a novel, metabolite-inspired chemical class capable of restoring chemotherapeutic sensitivity. The approach opens avenues for rational design of selective ABCB1 inhibitors and potential therapeutic applications in cancer, CNS drug delivery across the blood–brain barrier, HIV therapy, and potentially bacterial MDR ortholog targeting.

This study introduces a novel, bat-inspired strategy for ABCB1 inhibition using tryptophan-derived cyclic diketopiperazines. Through metabolomics, functional assays, computational modelling, and synthesis, the authors developed C3N-Dbn-Trp2, a benzylated Cyclo-(L-Trp-L-Trp) analogue that inhibits ABCB1 with IC50 values superior to verapamil in human cell models and effectively sensitizes drug-resistant cells to doxorubicin without impairing cell growth alone. The work establishes tryptophan-based scaffolds as viable leads and demonstrates how structural insights into ABCB1 binding can guide rational inhibitor optimization. Future research should focus on improving potency (approaching third-generation inhibitors), refining selectivity profiles, and undertaking comprehensive in vivo pharmacokinetic, safety, and efficacy studies, including assessment of impacts on normal ABCB1-expressing tissues and transporter stability dynamics in animal models.

- Potency gap: Although C3N-Dbn-Trp2 outperforms verapamil, it remains less potent than third-generation inhibitors like tariquidar (nanomolar range). - Safety not established in vivo: Lack of adverse effects in cell culture does not guarantee in vivo safety; comprehensive toxicity and PK studies are needed. - On-target effects in normal tissues: ABCB1 is expressed in select normal tissues; inhibition may impact physiological protective functions—this requires careful evaluation. - Stability and pharmacokinetics: In vivo stability (both instability and over-stability) could limit efficacy or increase side effects; needs assessment in animal models. - Partial off-target effects: Partial inhibition of ABCG2 observed at 10 µM necessitates further selectivity optimization.