← Back to Home

Finding a Partner for Drug-Resistant Prostate Cancer: AI Identifies Berberine Combination Clues

A UCLA team used language models to rank thousands of natural compounds, then tested them in three-dimensional cancer spheroids. Berberine and enzalutamide showed synergy in four models, but drug exposure and safety remain hurdles before use in humans.

By SURL BioNews

Prostate cancer cells can find alternative routes to survive treatment, an important reason drugs gradually lose effectiveness. Finding suitable partners for existing drugs might help block these escape routes. The MAESTRO platform, developed by a team at the University of California, Los Angeles (UCLA), brought AI-generated combination hypotheses into the laboratory and found synergy worth investigating between berberine and the prostate cancer drug enzalutamide in cell models.

The study was published in *Advanced Science* on October 8, and UCLA announced the findings on the 21st. Enzalutamide inhibits cancer cells by blocking androgen receptor signaling, but cancer cells may activate other pathways to escape. The study therefore addressed a specific question: Which compounds could disrupt these alternative pathways and make cancer cells respond to receptor blockade again?

MAESTRO used large language models to organize knowledge about compounds’ mechanisms of action, prostate cancer biology, and drug resistance, rank 3,592 natural products, and propose pathway explanations that could be tested experimentally. The team evaluated each candidate at least 100 times, comparing both average rank and ranking variability to select compounds that ranked highly and consistently. AI’s role here was to narrow the search; ranking consistency alone still cannot establish drug efficacy.

The researchers then paired 150 priority candidates with enzalutamide and tested them in three-dimensional prostate cancer spheroids. These models preserve some three-dimensional cellular interactions, but they are also prone to necrosis caused by oxygen deprivation at their centers, which can interfere with the interpretation of drug responses. The team therefore supplied additional oxygen to reduce cell death caused by culture conditions; the trade-off is that this environment also cannot fully reproduce the oxygen-deprived conditions of actual tumors.

Of the 150 compounds tested, 23 showed synergy with enzalutamide, meaning their combined effects exceeded what would be expected from their individual effects. Further testing showed that the berberine combination produced consistent synergy across four prostate cancer models with different mechanisms of drug resistance. MyNewsLA’s coverage of the same findings also reported these screening results, but these numbers represent laboratory hits and cannot be translated into patient treatment success rates.

Mechanistic analysis provided another layer of clues: Following combination treatment, activity of AMPK, which is involved in cellular energy and stress regulation, increased, while mTORC1 signaling, which promotes growth and survival, was inhibited. These changes were accompanied by reductions in DNA replication and cell cycle progression. The changes were consistent with the pathway hypotheses proposed by the model; however, the paper also noted that methods such as genetic manipulation and rescue experiments are still needed to clarify causal relationships among the signals.

The findings remain limited to in vitro research, with patient-derived tumor models and animal experiments still being planned. Berberine also faces the problem of low and highly variable systemic exposure after oral administration. Whether the concentrations required in the experiments can be achieved in human tumors, and whether the combination is safe, remain unanswered. This study offers a route from knowledge in the literature to testable drug treatment hypotheses, but does not yet provide sufficient support for patients to add berberine on their own.

References

  1. UCLA Health
  2. PubMed / Advanced Science
  3. MyNewsLA