Mechanism
Cancer adapted to a desert. Then you flood it.
Bipolar androgen therapy is a timed environmental shock. It is not testosterone replacement, and it is not a holiday from ADT.
Standard hormone therapy works at first because prostate-cancer cells are used to living in a testosterone-rich world. Strip the hormone away and many of them die. The ones that live rewrite themselves. They make far more androgen receptor — sometimes orders of magnitude more — so that a trickle of androgen is enough. That is castration resistance.
Samuel Denmeade and colleagues at Johns Hopkins treated that adaptation as a liability. If the cell has covered itself in receptor, a sudden surplus of ligand does not feed it. It jams it. Excess androgen-bound AR interferes with DNA relicensing in S phase, turns down growth programs (including MYC in some models), and can trigger stress and immune pathways. Cells stop dividing. Some die.
Flooding the cell with testosterone gums up the works. The repeated shocks do not give it time to adapt, because the environment is always changing.
Paraphrased from Denmeade’s patient guide and PCF interviews — not a verbatim clinical claim.
Why the swing has to be fast
Give a man a steady high dose and surviving clones will down-regulate AR and grow in the new normal. Leave him castrate and they will keep climbing AR. BAT tries to occupy neither steady state. Intramuscular testosterone cypionate 400 mg produces a high pole of roughly 1,000–3,000 ng/dL within a couple of days, then decays toward 100–200 ng/dL by day 28. The next injection restarts the cycle.
ADT does not stop. Lupron, Eligard, Orgovyx, Firmagon, or orchiectomy stay in place so testicular production cannot fill in the trough. If you stop ADT, you no longer have two poles. You have replacement.
Why PSA is a dishonest narrator
PSA is an androgen-receptor target gene. Raise testosterone and PSA can rise even when tumor volume is not. Denmeade describes three patterns: a true PSA response (drop of 50% or more that lasts), a plateau (higher but stable PSA with quiet scans and a man who feels better), and true progression (rising PSA plus growing disease). The plateau is the one that gets people taken off a working treatment.
Imaging decides. Bone scans can flare. PCWG3 rules exist for this. Early darker spots are not automatically new tumors.
The second gift: resensitization
BAT drives AR down. Enzalutamide and its cousins work better when AR is in play as a ligand-binding target. After a BAT course, those drugs often work again — sometimes better than they did the first time. TRANSFORMER’s crossover is the exhibit: 77.8% PSA50 for enzalutamide after BAT versus 23.4% for BAT after enzalutamide. STEP-UP is asking whether you can keep looping that sequence.
High baseline AR activity, AR amplification or mutation in blood, and in some series DNA-repair defects, are the biomarkers people are watching. None of them is a clinic-ready companion diagnostic yet.