Teaching Cancer Cells to Remember Who They Used to Be:

This image shows patient derived tumor organoids before (top) and after treatment (bottom). 
The colors show the activation of pathways related to cell differentiation in cancer stem cells. 
Shortly after these images were taken, the cancer stem cells spontaneously collapsed. 
Photo credit: Pradipta Ghosh/HUMANOID

What the New Science of Cancer Reversion Means for Prostate Cancer Patients

Breakthrough studies from Korea and San Diego have demonstrated that cancer cells can be reprogrammed back toward normality — not killed, but re-educated. The implications for prostate cancer patients — from newly diagnosed to late-stage mCRPC — are potentially profound.

 

Special Science Update — March 2026 Prepared by the IPCSG Educational Committee

 

Bottom Line Up Front A cluster of recent studies — led by researchers at KAIST in South Korea and UC San Diego — has demonstrated that cancer cells can be systematically converted back toward a normal state by flipping specific molecular switches, without killing the cells or editing their DNA. This approach, called

The Core Idea: Cancer as a Case of Forgotten Identity

For decades, cancer treatment has been organized around a single imperative: destroy the tumor. Chemotherapy, radiation, and even immunotherapy all work, in different ways, by eliminating malignant cells — often at considerable collateral cost to normal tissue.

A growing body of science now asks a different question. What if cancer cells are not destroyed beyond repair, but have simply lost the molecular instructions that tell them what they are supposed to be? Every cell in the body carries the complete human genome. What makes a prostate cell different from a liver cell is not the DNA sequence — it's the pattern of which genes are switched on and which are silenced, a system of molecular bookmarks collectively called the epigenome. Cancer, in this view, is partly a disease of the epigenome: the bookmarks get scrambled, the cell forgets its identity, and it starts behaving as an undifferentiated, proliferating cell with no specific function and no built-in limit on growth.

The exciting new proposition — now supported by preclinical evidence in colon cancer, and by proven clinical success in a leukemia subtype — is that if you can identify the key molecular switches that maintain the cancerous state, you may be able to flip them back, re-educating the cell rather than killing it..

Wait — Isn't Cancer Caused by DNA Mutations?

Yes — and that's still true. Cancer does involve real errors in the DNA sequence itself: mutations in genes like TP53, BRCA2, and PTEN that damage the cell's growth controls. Those errors are genuine and important.

But mutations alone don't fully explain cancer behavior. The same DNA sequence exists in every cell in your body — a liver cell and a prostate cell carry identical genetic text. What makes them different is the epigenome: a system of chemical tags and molecular "bookmarks" layered on top of the DNA that tells each cell which genes to read and which to ignore. Cancer disrupts both systems simultaneously. Mutations change some of the words in the text; epigenetic dysregulation scrambles the bookmark system, causing the cell to read the wrong chapters and ignore critical safety instructions.

Reversion therapy doesn't claim the mutations aren't there. It says that even a cell carrying mutations may retain enough epigenetic plasticity that restoring the right regulatory signals can override the consequences of those mutations — stopping uncontrolled growth and recovering something close to normal behavior. The DNA errors remain, but the cell's identity programming is strong enough to reassert itself. The APL leukemia cure is the proof: those cells carry a real chromosomal abnormality, yet ATRA forces them to mature and function normally anyway, because the epigenetic block on normal development is removed.

Think of it this way: the mutations change a few words in the book. Epigenetic dysregulation causes the cell to lose its place entirely and start reading from a chaotic, dangerous chapter. Reversion therapy tries to find the page again.

We Know This Can Work: The Clinical Proof and the New Evidence

Cancer reversion is not theoretical. Since the late 1980s, a form of leukemia called acute promyelocytic leukemia (APL) — once one of the deadliest blood cancers — has been treated with a vitamin A derivative (all-trans retinoic acid, or ATRA) that forces stalled leukemia cells to complete their maturation into normal blood cells. Combined with arsenic trioxide, this approach now achieves cure rates approaching 90%, entirely through re-education rather than destruction. APL is now the most curable adult leukemia.

The challenge has been extending this principle to solid tumors, where the biology is far more complex. That challenge is now being addressed with impressive precision.

In December 2024, researchers at KAIST (Korea Advanced Institute of Science and Technology) published a landmark study in Advanced Science showing that colon cancer cells could be systematically converted back to normal intestinal cells by simultaneously silencing just three master regulatory genes — MYB, HDAC2, and FOXA2. The team used a computational "digital twin" model of normal colon cell development to identify these switches, then validated the finding in multiple human cancer cell lines and in mice. The reprogrammed cancer cells slowed their growth, lost their aggressive behavior, and molecularly resembled healthy intestinal tissue. Crucially, no DNA editing was performed — the cells were guided back through changes in gene expression alone.

A follow-up KAIST paper published in January 2025 went further, identifying the precise molecular "tipping point" at which normal cells tip over into cancer — and showing that a reversion switch can be found and applied at that vulnerable transitional moment, before the malignant state is fully consolidated.

In October 2025, UC San Diego researchers published a complementary study in Cell Reports Medicine, using an AI tool called CANDiT to identify treatments that reprogram the most dangerous and drug-resistant colon cancer cells — cancer stem cells — back toward normality. Remarkably, once reprogrammed to behave like normal cells, those cancer stem cells spontaneously self-destructed. Modeling across more than 2,100 patient datasets suggested this approach could reduce recurrence and death risk by up to 50%. A clinical-grade drug candidate already exists and has passed Phase 1 safety testing.

What "Reversion" Actually Means at the Cell Level

Normal cells carry an identity code in their epigenome — a pattern of molecular tags on DNA and the proteins that package it, determining which genes are active. When cells become cancerous, key parts of this code are overwritten: genes that promote growth become permanently switched on; genes that enforce normal function and limit division are silenced. Reversion therapy aims to restore critical parts of the original code — not by editing DNA, but by inhibiting the specific regulatory proteins that maintain the cancerous pattern. Once enough of the identity code is restored, the cell's own internal logic takes over and drives it back toward normal behavior.

Why This Is Directly Relevant to Prostate Cancer

The KAIST and UCSD studies were conducted in colon cancer models. But prostate cancer patients should not file this under "interesting but not my disease." The molecular biology of prostate cancer progression is, in important respects, a textbook example of exactly the kind of cellular identity loss that reversion therapy targets.

Prostate Cancer is Already a Disease of Misremembered Identity

Normal prostate cells depend on androgen receptor (AR) signaling for their identity — it is the master regulator that tells them to be prostate cells and to behave like prostate cells. When prostate cancer progresses, the AR network is progressively hijacked and reprogrammed. Research published in Nature Genetics showed that in metastatic castration-resistant prostate cancer (mCRPC), the AR is literally rerouted to regulatory sites associated with fetal prostate development — the cancer cell is not just growing out of control, it is reaching backward in developmental time to reactivate programs from before the tissue was fully formed. Each major treatment advance — surgical castration, medical castration, enzalutamide, abiraterone — puts pressure on cancer cells that drives further identity distortion.

The most dangerous endpoint of this distortion is neuroendocrine prostate cancer (NEPC) — a transformation in which prostate cancer cells abandon their prostatic identity almost entirely and acquire features of nerve cells, becoming completely AR-independent and resistant to essentially all current hormonal therapies. Approximately 30% of patients with mCRPC eventually progress to NEPC or mixed neuroendocrine features. The molecular drivers of this transformation — EZH2, SOX2, FOXA2, N-MYC — are epigenetic regulators of the same general class as the switches identified by the KAIST team. In fact, FOXA2, one of the three KAIST "reversion switches" in colon cancer, is also a key driver of NEPC development in prostate cancer. The biology is intimately connected.

HDAC Inhibitors and EZH2 Inhibitors Are Already Being Tested in Prostate Cancer

Several of the drug classes most directly relevant to reversion therapy have already entered prostate cancer clinical trials. HDAC inhibitors — which loosen the tightly packed DNA that silences tumor-suppressor genes — are being studied in combination with AR inhibitors in mCRPC. EZH2 inhibitors, which block a key epigenetic regulator that drives both castration resistance and NEPC progression, are the subject of multiple ongoing trials (including NCT03460977 and NCT04846478 among others). Importantly, preclinical research has shown that EZH2 inhibition can reverse the NEPC phenotype and restore sensitivity to enzalutamide — a direct example of reversion logic applied to prostate cancer. BET inhibitors (targeting BRD4, which drives AR signaling in CRPC) have also shown preclinical efficacy and are in clinical development.

None of these are yet standard of care, and their clinical results so far have been mixed — reflecting the challenge that a single epigenetic intervention often provides incomplete reversion, and combination strategies will likely be needed. But the principle is being actively tested.

"Nipping It in the Bud": The Early-Stage Opportunity

One of the most intriguing questions raised by the KAIST REVERT study is whether reversion therapy could one day be applied very early — not after cancer is fully established, but at the transition point when cells are just beginning to lose their normal identity.

This is not science fiction. The KAIST team specifically showed that the molecular transition from normal to cancerous is not abrupt — there is a window of cellular instability during which the cell has not yet fully committed to malignancy. At this point, relatively modest molecular interventions may be sufficient to tip the balance back toward normal, because the cancerous state is not yet strongly stabilized.

For prostate cancer patients, this has tantalizing implications. Men on active surveillance for low-grade, localized prostate cancer (Gleason 6 / Grade Group 1) are, by definition, living with cancer cells that have not yet acquired the full epigenetic profile of aggressive disease. If the molecular transition toward higher-grade disease could be intercepted and reversed — not with surgery or radiation, but with a targeted epigenetic agent — the implications for the active surveillance population would be enormous. The goal would shift from "watching and waiting to see if we need to treat" to "actively maintaining cellular normalcy."

Similarly, in the months after radical prostatectomy or radiation, when PSA is undetectable but disseminated tumor cells may already be present in bone marrow and lymph nodes in a dormant state, early reversion or dormancy-reinforcement strategies could potentially prevent those cells from ever completing the transition to active metastatic disease.

The Dormancy Connection: A Clinical Trial Already Underway

Research at institutions including Albert Einstein College of Medicine has identified a molecular pathway that keeps disseminated tumor cells — the individual cells that escape the primary tumor and travel to distant sites — in a dormant, non-proliferating state. A key player is a nuclear receptor called NR2F1, which is normally silenced in tumors but is spontaneously upregulated in dormant disseminated cells. Retinoic acid (related to the ATRA used in APL treatment) drives NR2F1 expression and enforces dormancy.

A combination of low-dose 5-azacytidine (AZA, a DNA demethylating agent) followed by all-trans retinoic acid (atRA) has been shown to restore NR2F1-driven dormancy in disseminated tumor cells and suppress their outgrowth into overt metastases in preclinical models. This AZA+atRA strategy is currently being tested in a clinical trial in prostate cancer patients — making it one of the first direct clinical applications of dormancy-reinforcement/reversion therapy specifically in prostate cancer.

The Late-Stage Problem: Micrometastases, Drug Resistance, and Identity Collapse

For men with mCRPC, the clinical situation is defined by a set of deeply frustrating realities. Current imaging can detect bone lesions, lymph node involvement, and visceral metastases — but it cannot see individual cancer cells or tiny clusters of cells, the micrometastases that represent the true seeds of disease progression. Lutetium-177 PSMA therapy has shown remarkable efficacy precisely because it can deliver radiation at the cellular level, and Phase 1/2 data on terbium-161 PSMA therapy (presented at ASCO 2025) suggests even finer-grained targeting may be achievable. But radioligand therapies require cells to express PSMA — and as disease progresses toward NEPC, PSMA expression often drops or disappears, leaving those cells invisible and untargetable.

This is exactly where reversion therapy becomes most compelling as a complementary strategy. The cancer cells that are hardest to find and kill with current tools — the PSMA-low, AR-independent, neuroendocrine-like cells that represent the leading edge of mCRPC progression — are also the cells that have undergone the most dramatic identity loss. They have, in a sense, transformed themselves most completely from what they originally were. The question reversion science asks is: could they be transformed back?

Re-sensitizing Resistant Cells

Here the evidence is genuinely encouraging. Multiple preclinical studies have shown that epigenetic interventions can reverse NEPC-like features and restore AR expression — essentially making cells hormone-sensitive again after they had become hormone-independent. EZH2 inhibition, in particular, has been shown to reverse NEPC differentiation and re-sensitize cells to enzalutamide in laboratory models. If this could be achieved clinically, it would represent a way to "reset" a patient's disease to a state where existing hormonal therapies work again.

The UC San Diego CANDiT study adds another angle: the observation that forcing cancer stem cells back toward a differentiated state caused them to self-destruct. If something analogous could be demonstrated in prostate cancer stem cells — which express their own set of stemness markers, including CD44, and are believed to be a primary driver of resistance and recurrence — the therapeutic implications would be significant. Cells that resist every conventional therapy because they are defined by their stem-like identity might be uniquely vulnerable to losing that identity.

Early/Localized Disease Opportunity

Target: Men on active surveillance, or post-treatment with undetectable PSA but micrometastatic risk.

Goal: Intercept the transition toward higher-grade disease; enforce dormancy of any disseminated cells.

Current status: AZA+atRA clinical trial active in prostate cancer. EZH2/HDAC inhibitor trials ongoing. Computational frameworks (BENEIN, REVERT, CANDiT) being extended to prostate cancer biology.

Late-Stage / mCRPC Opportunity

Target: PSMA-low/AR-independent/neuroendocrine-like cells; drug-resistant cancer stem cells; micrometastases below imaging threshold.

Goal: Reverse identity loss; re-sensitize cells to existing therapies; eliminate cells that cannot survive without their cancerous identity.

Current status: EZH2/BET/HDAC inhibitor trials in mCRPC. Preclinical reversal of NEPC demonstrated. No approved reversion therapy for prostate cancer yet.

The Road Ahead: What Has to Be Solved

Translating these findings into prostate cancer treatments will require several things to happen. First, the computational frameworks developed for colon cancer — BENEIN, REVERT, CANDiT — need to be systematically applied to prostate cancer single-cell data to identify the prostate-specific master regulatory switches. The human prostate cancer single-cell atlas is developing rapidly, providing the raw material for this analysis. Second, drug delivery to micrometastatic cells in bone and lymph nodes presents different challenges from treating a colon tumor, and will require advances in targeted delivery. Third, the stability problem must be addressed: the December 2025 comprehensive review of cancer reversion therapy noted that while APL-type reversion produces durable normalization, some forms of epigenetic reversion are stimulus-dependent and reverse when treatment stops — designing approaches that produce lasting cellular change rather than temporary suppression is essential.

Combination approaches are likely to be key. Reversion therapy is probably most powerful not as a standalone replacement for current treatments, but as an addition that addresses the cells current therapies cannot reach: the dormant disseminated cell, the stem-like treatment-resistant subpopulation, the NEPC-transitioning cell at the edge of the tumor. Used in combination with PSMA-targeted radioligand therapy, checkpoint inhibitors, or AR inhibitors, reversion agents could potentially close the gaps that currently allow fatal disease progression.

What This Means for You Right Now

None of the colon cancer reversion approaches (BENEIN/REVERT/CANDiT) have been tested in human clinical trials. They remain preclinical. The prostate cancer-specific epigenetic trials (EZH2 inhibitors, HDAC inhibitors, AZA+atRA) are in Phase 1/2 and not yet standard of care.

Patients should not attempt to use HDAC inhibitors or other epigenetic drugs off-label based on this research — these agents have real toxicities and require careful clinical management.

What you can do: Ask your oncologist whether any epigenetic combination trials are open at your institution or at a nearby academic center. For men with mCRPC who are progressing through hormonal therapies, participation in trials combining epigenetic agents with standard therapies may become an increasingly attractive option.

IPCSG will continue to track these developments and report on clinical trials as they open.

Conclusion: A Smarter Way to Think About What Cancer Is

The science described here represents more than a collection of interesting laboratory results. It represents a shift in how we conceptualize cancer itself - a new paradigm — from a permanent, irreversible corruption of cellular machinery to a set of regulatory mistakes that, in at least some circumstances, can be corrected. The proof of concept exists in APL. The computational tools to extend it systematically to other cancers now exist and are being applied. Clinical trials in prostate cancer are already underway at the early end of this paradigm.

For prostate cancer patients, this matters at both ends of the disease spectrum. For men with early disease, it raises the possibility that future treatment might one day focus on maintaining cellular identity rather than destroying cells. For men with late-stage disease, it offers the prospect of addressing the cells that have escaped every conventional weapon in our arsenal — not by finding a more powerful way to kill them, but by reminding them what they used to be.

As Professor Kwang-Hyun Cho of KAIST put it: "The fact that cancer cells can be converted back to normal cells is an astonishing phenomenon. This study proves that such reversion can be systematically induced." The word systematically is the key advance. We are no longer waiting for happy accidents. We are learning to engineer the cellular memory of cancer — and for prostate cancer patients, that is a development worth watching closely.

Verified Sources and Formal Citations

1. Gong, J.-R., Lee, C.-K., Kim, H.-M., et al. (2024). Control of Cellular Differentiation Trajectories for Cancer Reversion. Advanced Science, 12(3), e2402132. Published December 11, 2024. DOI: 10.1002/advs.202402132 | PubMed: PMID 39661721
2. Shin, D., Gong, J.-R., Jeong, S.D., et al. (2025). Attractor Landscape Analysis Reveals a Reversion Switch in the Transition of Colorectal Tumorigenesis. Advanced Science, 12(8), e2412503. Published January 22, 2025. DOI: 10.1002/advs.202412503 | PubMed: PMID 39840939
3. Sinha, S., Alcantara, J., Perry, K., et al. (2025). CANDiT: A machine learning framework for differentiation therapy in colorectal cancer. Cell Reports Medicine, 6(11), 102421. Published October 20, 2025. DOI: 10.1016/j.xcrm.2025.102421 | PubMed: PMID 41118768
4. KAIST News Center. (December 23, 2024). Foundational Technology to Make Cancer Cells Revert to Normal Cells. https://news.kaist.ac.kr/newsen/html/news/?mng_no=42710&mode=V
5. KAIST News Center. (February 5, 2025). Discovery of Molecular Switch That Reverses Cancerous Transformation at the Critical Moment of Transition. https://news.kaist.ac.kr/newsen/html/news/?mode=V&mng_no=43810
6. UC San Diego Health. (October 16, 2025). Precision Reprogramming: How AI Tricks Cancer's Toughest Cells. EurekAlert! https://www.eurekalert.org/news-releases/1102117
7. [Cancer Reversion Review] Pensotti, A. & Bizzarri, M. (2024). The phenotypic reversion of cancer: Experimental evidences on cancer reversibility through epigenetic mechanisms. Oncology Reports, 51(3), 48. PMC: PMC10835663
8. [Cancer Reversion Therapy Review] Cancers (MDPI). (December 15, 2025). Cancer Reversion Therapy: Prospects, Progress and Future Directions. https://www.mdpi.com/1467-3045/47/12/1049
9. [APL Differentiation Therapy] Thomas, X. (2019). Acute Promyelocytic Leukemia: A History over 60 Years. Oncology & Therapy, 7(1), 33–65. PubMed: PMID 32700196
10. [Frontiers in Oncology] (December 28, 2022). The treatment of acute promyelocytic leukemia in 2023: Paradigm, advances, and future directions. https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.1062524/full
11. [Prostate cancer epigenomics / metastasis] Pomerantz, M.M., et al. (2020). Prostate cancer reactivates developmental epigenomic programs during metastatic progression. Nature Genetics, 52, 790–799. PubMed: PMID 32690948
12. [NEPC and epigenetic reprogramming] Blee, A.M. & Huang, H. (2022). The Role of Epigenetic Change in Therapy-Induced Neuroendocrine Prostate Cancer Lineage Plasticity. Frontiers in Endocrinology, 13. PMC: PMC9329809
13. [NEPC treatment targets] Kwan, E.M., et al. (2024). New Insights into Potential Therapeutic Targets for Neuroendocrine Prostate Cancer: From Bench to Clinic. Research (Science Partner Journal). https://spj.science.org/doi/10.34133/research.0791
14. [Epigenetic (de)regulation in prostate cancer] Li, Y. & Dahiya, R. (2024). Epigenetic (de)regulation in Prostate Cancer. Cancer Treatment and Research. PMC: PMC11421856
15. [Epigenetic subtype heterogeneity, NEPC, EZH2 in CRPC] Epigenetic mechanisms underlying subtype heterogeneity and tumor recurrence in prostate cancer. Nature Communications (2023). https://www.nature.com/articles/s41467-023-36253-1
16. [Epigenetic reprogramming in cancer stemness and resistance] Discover Oncology (2025, November 16). Epigenetic reprogramming as the nexus of cancer stemness and therapy resistance. https://link.springer.com/article/10.1007/s12672-025-04085-8
17. [Dormancy and AZA+atRA in prostate cancer] Albert Einstein College of Medicine Cell Biology Department overview (2024–2025), citing active prostate cancer clinical trial of AZA+atRA strategy for disseminated cancer cell dormancy. https://einsteinmed.edu/uploadedFiles/education/phd/cell-biology-orientation-booklet-2024-2025.pdf
18. [Cancer dormancy mechanisms and micrometastases] Towards understanding cancer dormancy over strategic hitching up mechanisms to technologies. Molecular Cancer (2025, February 14). https://molecular-cancer.biomedcentral.com/articles/10.1186/s12943-025-02250-9
19. [AACR Special Conference] Cancer Dormancy and Residual Disease. American Association for Cancer Research. https://www.aacr.org/meeting/aacr-special-conference-in-cancer-research-cancer-dormancy-and-residual-disease/
20. [Prostate cancer ASCO 2025 highlights, terbium-161 for micrometastases] Medscape. (June 24, 2025). Prostate Cancer Highlights From ASCO 2025. https://www.medscape.com/viewarticle/prostate-cancer-updates-treatment-asco-2025-2025a100010w

Informed Prostate Cancer Support Group (IPCSG)
This newsletter article is for educational purposes only and does not constitute medical advice.
Please consult your oncologist before making any changes to your treatment plan.