Old Antibiotics, New Trick: Doxycycline Blocks Prostate Cancer Spread to Bone

One of the most feared turning points in prostate cancer is the moment the disease reaches the skeleton. Bone metastasis transforms a cancer that is often manageable for years into a painful, largely incurable condition, and it remains a common and lethal complication not only of prostate cancer but also of breast and lung cancers. Now, a team at Southern University of Science and Technology in Shenzhen reports that two familiar antibiotics, doxycycline and methacycline, can shut down a molecular switch that drives this spread, offering a striking example of drug repurposing guided by a precise mechanistic target.

The switch in question is a chemically modified version of a transcription factor called KLF5, or Krüppel-like factor 5. In earlier work, the same group showed that transforming growth factor beta, or TGF-β, a signaling molecule abundant in the bone microenvironment, induces the addition of an acetyl group to KLF5 at a specific position, lysine 369. This acetylated form of KLF5, abbreviated Ac-KLF5, acts as a master regulator of malignancy: it drives epithelial-mesenchymal transition, the process by which cancer cells shed their adhesive character and become invasive, and it promotes the osteolytic destruction of bone that characterizes metastatic lesions. Because Ac-KLF5 sits at the convergence of bone-derived signals and metastatic behavior, the researchers reasoned that it, and the genes it controls, would make an ideal therapeutic target.

Finding a drug that disables a transcription factor is notoriously difficult, since these proteins lack the deep pockets that small molecules typically exploit. The team therefore turned to an unbiased, systematic approach. Using surface plasmon resonance, a label-free biophysical technique that measures binding in real time, they screened a library of 1,987 FDA-approved compounds against a purified mimic of acetylated KLF5, a mutant in which lysine 369 is replaced by glutamine, known as KQ. This substitution mimics the charge and structural consequences of acetylation, allowing the researchers to fish for molecules that recognize the acetylated state specifically.

The screen produced a clear and surprising winner: the tetracycline family of antibiotics. Doxycycline and methacycline both bound selectively to the KQ protein, but crucially, they showed no detectable binding to an acetylation-deficient mutant, KLF5 K369R, in which lysine is replaced by arginine. The parent compound, tetracycline itself, failed to bind either protein. The selectivity was verified not only by surface plasmon resonance but also by circular dichroism spectroscopy, which tracks changes in protein secondary structure upon ligand binding. In other words, the anti-metastatic potential of these old antibiotics appears to hinge on a chemical modification of KLF5 that ordinary tetracycline cannot recognize.

Binding alone proves nothing unless it changes cell behavior, so the team moved to functional assays. In prostate cancer cells engineered to express the acetylation-mimicking KQ mutant, doxycycline and methacycline significantly inhibited invasion in three-dimensional spheroid invasion assays and Transwell migration assays, while tetracycline again had no effect. The drugs also suppressed the ability of cancer cells to promote osteoclast differentiation, the process by which bone-resorbing cells are recruited and activated in metastatic lesions. These results established that the physical interaction between the antibiotics and acetylated KLF5 translates directly into reduced metastatic behavior in vitro.

To test therapeutic efficacy in a living system, the researchers used tibial injection xenograft models, in which human prostate cancer cells are introduced directly into the mouse tibia, recreating the osteolytic bone metastasis that plagues patients. Tumors were tracked with bioluminescence imaging, bone destruction was quantified by micro-computed tomography, and lesions were examined histologically. Doxycycline and methacycline treatment markedly reduced tumor burden and osteolytic damage in this setting, whereas tetracycline again failed to help. The in vivo data, generated under approved animal ethics protocols at the Shenzhen institution, provided the strongest evidence yet that targeting Ac-KLF5 can blunt bone metastasis in a whole organism.

The mechanistic core of the study came from integrating two genome-wide datasets. RNA sequencing revealed which genes change their expression when the drugs are applied, while chromatin immunoprecipitation sequencing, or ChIP-seq, mapped where acetylated KLF5 binds directly to gene promoters. The overlap yielded 22 genes whose transcription is regulated both by KQ through direct promoter binding and by doxycycline and methacycline treatment, identifying a compact drug-sensitive transcriptional program downstream of Ac-KLF5. Quantitative PCR, small interfering RNA knockdown, and rescue experiments validated these connections.

Among the 22 genes, one stood out: PGK1, which encodes phosphoglycerate kinase 1, a glycolytic enzyme best known for its role in energy metabolism but increasingly recognized as a contributor to cancer progression. Follow-up experiments confirmed that PGK1 is responsible for the suppressive effects of doxycycline and methacycline on cell invasion, osteoclast differentiation, and bone metastasis. The finding ties metastatic aggression to metabolic rewiring and suggests that acetylated KLF5 acts partly by elevating a glycolytic enzyme that fuels invasive behavior and supports the bone-destroying cycle of metastasis.

The clinical implications are considerable. Doxycycline is cheap, orally available, and carries decades of safety data from its use as an antibiotic and anti-inflammatory agent, which means a repurposing pathway could move faster than de novo drug development. The study also resolves a long-standing puzzle: tetracycline derivatives were known to show anti-metastatic activity, but their molecular targets had remained unidentified, limiting rational clinical use. By pinning that activity to selective binding of acetylated KLF5 and its control of PGK1, the researchers have converted an empirical observation into a targetable mechanism, and their screening strategy offers a template for finding inhibitors of other acetylated transcription factors.

Cautions remain before patients benefit. The work relied on cell models and mouse xenografts rather than human trials, and the drug concentrations and schedules used experimentally may not map directly onto clinical dosing. Castration-resistant and metastatic castration-resistant prostate cancer are treated today with docetaxel and androgen-targeting agents, and any antibiotic-based regimen would need to demonstrate added benefit in combination or as maintenance therapy. Nonetheless, the identification of acetylated KLF5 as a druggable node in TGF-β-driven bone metastasis, published open access in BMC Medicine by Duo Zhang, Xinyuan Jiang, Yutian Ren and colleagues under the correspondence of Jin-Tang Dong, gives researchers a concrete molecular bullseye in one of oncology’s most lethal scenarios, and it suggests that a fifty-year-old antibiotic may yet have a role to play in keeping prostate cancer out of the bone.

Subject of Research: Targeting acetylated KLF5 with tetracycline antibiotics to inhibit prostate cancer bone metastasis

Article Title: Acetylated KLF5 as a therapeutic target mediating the anti-metastatic effects of doxycycline and methacycline in prostate cancer bone metastasis

Article References: Zhang, D., Jiang, X., Ren, Y., Qin, Y., Lin, X., Deng, S., Zhou, Q., Chen, Y., Zhang, Z., Zhang, X., & Dong, J.-T. (2026). Acetylated KLF5 as a therapeutic target mediating the anti-metastatic effects of doxycycline and methacycline in prostate cancer bone metastasis. BMC Medicine. https://doi.org/10.1186/s12916-026-05190-3

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05190-3

Keywords: acetylated KLF5, prostate cancer, bone metastasis, doxycycline, methacycline, tetracycline derivatives, TGF-beta, drug repurposing, PGK1, epithelial-mesenchymal transition, osteoclast differentiation, BMC Medicine