A large screen identifies beta-lactam antibiotics which can be repurposed to target the syphilis agent

Syphilis is an increasing global public health threat with millions of new cases annually and significant rises in congenital syphilis. Clinical manifestations are diverse, and untreated infection can progress to severe, irreversible disease. Understanding of Treponema pallidum has been limited historically by the lack of a robust in vitro cultivation system; however, long-term in vitro growth has recently been achieved, enabling new research directions. Current first-line therapy is a single intramuscular dose of benzathine penicillin G, but 1–10% of the population reports penicillin allergy, supply shortages persist, and prior use of alternative antibiotics (e.g., macrolides) has led to resistance. Doxycycline is the only recommended alternative but is limited by dosing duration, stage-specific efficacy, and contraindication in pregnancy. The authors hypothesized that additional β-lactams could be equally or more effective against T. pallidum and that continuous in vitro culture would enable systematic screening. The study aims to screen a large panel of β-lactams, identify top candidates, quantify their potency, and probe their effects on peptidoglycan synthesis.

Background literature highlights: (1) Historical difficulties and recent breakthroughs in continuous in vitro cultivation of T. pallidum enabled sustained propagation without loss of viability or infectivity; (2) Benzathine penicillin G has been standard therapy due to safety and efficacy, though allergy, shortages, and resistance concerns necessitate alternatives; (3) Prior macrolide resistance has been documented; (4) β-lactam action depends on outer membrane permeability and binding to essential penicillin-binding proteins (PBPs). In T. pallidum, PBPs have been studied, including reports of atypical β-lactamase activity and peptidoglycan-related functions; (5) Ceftriaxone has been previously assessed against T. pallidum with supportive clinical and pharmacokinetic data, including CNS penetration. The study builds on these findings by performing the first large-scale, systematic β-lactam screen using continuous in vitro culture and by integrating molecular readouts and peptidoglycan labeling.

Organism and co-culture: Treponema pallidum subsp. pallidum, strain Nichols, co-cultured with rabbit epithelial cells (Sf1Ep/SF1p) under microaerobic conditions (1.5% O2, 5% CO2, 93.5% N2). Cells maintained in MEM-based media for Sf1Ep and TpCM2 for T. pallidum, following published protocols. Cultures were passaged weekly; bacterial densities determined by microscopic enumeration using a Fuchs-Rosenthal hemocytometer. Growth kinetics and morphology: Continuous culture monitored to generate sawtooth growth curves; mean doubling time calculated. Cell length distributions assessed from phase-contrast microscopy of fixed cells. Automated cell detection performed using software (Outif/QUpid) with custom parameterization; >400 cells analyzed to fit length distributions to Gaussian models. High-throughput antibiotic screen: An 89-compound β-lactam library (Selleckchem nL300) plus tetracyclines (e.g., doxycycline) screened at 5 nM. Compounds solubilized in water or DMSO; corresponding solvent controls included. T. pallidum inoculated at 1×10^6 cells per well in 12-well plates with Sf1Ep monolayers. After 7 days, cells were harvested, pooled with rinses, diluted, and enumerated microscopically. Relative efficacy defined as percent cells remaining versus solvent control. Generational analyses performed for penicillin and cephalosporin subclasses. Top 25% performers re-evaluated in independent cultures/batches. Molecular efficacy assessment: RNA extracted after 1-week treatments from co-cultures; Sf1Ep monoculture RNA used as negative control. DNase treatment applied. One-step RT-qPCR targeting tp47 (penicillin-binding protein gene) and flaA (flagellar sheath protein) with validated primers. Genomic DNA standards used for absolute quantification. Melting curve analysis and no-template controls included. MIC determination: For top candidates (six from qRT-PCR plus ceftazidime and cefmenoxime), serial titrations (0.125–8 nM) in duplicate biological replicates performed. MIC defined as concentration yielding culture density ≤ inoculum. Modified Gompertz regression used to interpolate MIC values; means and SDs reported. Peptidoglycan synthesis labeling (HADA): Cultures pulse-labeled with 0.5 mM 7-hydroxycoumarin-3-carboxylic acid–D-alanine (HADA) for 24 h; fixed and imaged (phase-contrast and epifluorescence). For drug impact on PG synthesis, cultures treated with antibiotics at 2× MIC for 24 h, then HADA-labeled. To avoid cytoplasmic accumulation confounds, purified sacculi were prepared by boiling in 10% SDS, washing, and analyzing HADA fluorescence in isolated peptidoglycan. Demograph analysis conducted to assess spatiotemporal PG incorporation across cell-length–ordered populations. Microscopy and image analysis: Zeiss Axio Observer with oil-immersion objective and Hamamatsu Orca-Flash camera; DAPI channel for HADA fluorescence. Standardized acquisition; QUpid and custom MATLAB scripts used for high-throughput quantitative analysis. Data availability: All data and analysis scripts in figures/supplementary materials.

- Continuous in vitro cultivation achieved stable growth: mean doubling time 52.3 ± 13.2 hours; >10 cumulative generations; sustained growth for over a year. Population mean cell length was 13.31 ± 3.15 µm with a near-Gaussian distribution (R² = 0.8556) and slight skew toward longer cells. - At a screening concentration of 5 nM, penicillins were slightly more effective than cephalosporins overall, and both outperformed tetracyclines (e.g., doxycycline was ineffective at 5 nM). - Among penicillins, first- and fourth-generation agents significantly outperformed second- and third-generation agents. Cephalosporin generations showed no clear class-wide differences, though several were effective. - Top 25% of compounds were re-tested and evaluated molecularly by RT-qPCR of tp47 and flaA; transcript measures correlated well and enabled higher-resolution discrimination. Benzathine penicillin G produced no detectable transcripts, confirming strong activity. Six β-lactams emerged as top performers for further MIC testing. - MICs for top candidates were in the low nanomolar range: azlocillin sodium 0.530 nM; mezlocillin sodium 0.723 nM; benzathine penicillin G 1.091 nM; cefmetazole sodium 1.160 nM; nafcillin sodium 1.216 nM; ceftriaxone sodium 3.296 nM; cefazolin 3.750 nM; cefmenoxime hydrochloride 5.047 nM. Azlocillin and mezlocillin were approximately 15–20 times more potent than benzathine penicillin G by MIC. All candidates exhibited MICs 10–100 times lower than doxycycline. - HADA labeling revealed near-ubiquitous lateral peptidoglycan incorporation during most of the cell cycle with increased mid-cell signal late in the cycle (septal growth). Drug-treated purified sacculi showed significantly reduced HADA incorporation, confirming inhibition of peptidoglycan synthesis at the cell wall level. - Additional practical attributes: ceftriaxone demonstrated favorable half-life and CNS penetration; some penicillins (e.g., nafcillin) may offer availability and cost advantages, whereas azlocillin and mezlocillin are not currently produced in the U.S.

The study directly addresses the need for alternative syphilis therapeutics by systematically screening β-lactams against T. pallidum using continuous in vitro culture. Results show that numerous β-lactams can inhibit growth at low nanomolar concentrations, with several outperforming benzathine penicillin G in vitro. These findings are significant given ongoing benzathine penicillin shortages, variable allergy profiles, and prior resistance to non-β-lactam alternatives. The identification of azlocillin and mezlocillin as most potent by MIC, and the strong activity of other penicillins (e.g., nafcillin) and cephalosporins (e.g., ceftriaxone, cefmetazole, cefazolin, cefmenoxime), expands the candidate pool for future in vivo testing and clinical consideration. Pharmacokinetic and clinical attributes (e.g., ceftriaxone’s long half-life and CNS penetration) suggest potential utility in neurosyphilis or late-stage disease. Cross-reactivity patterns in β-lactam allergy indicate that some candidates may be tolerated in penicillin-sensitive patients. Beyond therapeutics, the HADA labeling and demograph analyses provide insights into T. pallidum cell biology, indicating predominantly lateral peptidoglycan synthesis with septal increase late in the cycle, and sacculi properties distinct from Borrelia, adding fundamental knowledge that could inform mechanism-targeted interventions. Availability and cost considerations, as well as differences in structural R-groups that may affect allergy risk, are discussed in interpreting the translational potential of candidates.

This work presents the first large-scale in vitro screen of nearly 100 β-lactam antibiotics against Treponema pallidum, enabled by continuous culture, and identifies multiple candidates with low-nanomolar MICs. Azlocillin and mezlocillin exhibited the lowest MICs, outperforming benzathine penicillin G, while several cephalosporins, including ceftriaxone, also demonstrated strong activity. Molecular and peptidoglycan-labeling assays corroborated growth inhibition and mechanistic effects on cell wall synthesis. The study expands the repertoire of potential syphilis therapies and introduces quantitative cell-biological tools for T. pallidum research. Future work should include in vivo validation, pharmacokinetic/pharmacodynamic optimization, assessment across disease stages (including CNS involvement), evaluation in penicillin-allergic populations, and clinical trials to determine efficacy, dosing, and safety.

- In vitro study using a single T. pallidum strain (Nichols) co-cultured with rabbit epithelial cells; generalizability to other strains and in vivo conditions is uncertain. - Initial screening at a single concentration (5 nM) may under- or overestimate activity for some agents; although MIC titrations were performed for a subset, not all screened compounds were titrated. - Microscopic enumeration has a detection limit (~1000 cells/mL) and relies on visible spirochetes, potentially missing subpopulations. - qRT-PCR targets (tp47, flaA) may not capture all mechanisms of growth inhibition. - HADA labeling required sacculi purification to avoid cytoplasmic uptake artifacts; while addressed, residual confounds are possible. - Translational factors such as drug availability (e.g., azlocillin/mezlocillin not produced in the U.S.), dosing regimens, route of administration, tissue penetration (including CNS), allergy cross-reactivity, and real-world costs were not experimentally evaluated. - No animal or clinical efficacy data are provided; results require in vivo validation.