Cancer Stem Cells: The Hidden Driver of Metastasis, Relapse and Drug Resistance (2026)
Abstract
Cancer stem cells (CSCs) are cancer-cell populations or cellular states characterized by properties such as self-renewal, tumor initiation, differentiation capacity, metabolic adaptation and resistance to therapy. CSC biology has become increasingly important in research on cancer recurrence, metastasis and treatment resistance.
However, the traditional model of CSCs as a rare, permanently defined population is being revised. Contemporary evidence supports a more dynamic model in which cancer cells can transition between stem-like and more differentiated states. Tumor microenvironmental signals, hypoxia, epithelial-mesenchymal plasticity, epigenetic regulation and metabolic changes can influence this process.
Potential CSC-directed strategies include inhibition of Wnt/β-catenin, Hedgehog, Notch, PI3K/AKT/mTOR, JAK/STAT3, NF-κB, TGF-β and YAP/TAZ signaling, as well as targeting markers such as ALDH, CD44 and CD133. Several approved drugs affect pathways implicated in CSC biology, while many other proposed CSC inhibitors remain experimental.
This review examines how common CSCs may be, why CSC frequency is difficult to quantify, the major therapeutic targets, the evidence surrounding repurposed drugs, and why future CSC therapy will probably require combinations targeting both tumor drivers and cellular plasticity.
1. What Are Cancer Stem Cells?
Cancer stem cells are a functional concept rather than a single universally identifiable cell type. CSCs are generally characterized by their ability to self-renew, generate heterogeneous tumor-cell populations and, under appropriate experimental conditions, initiate or sustain tumor growth.
The original CSC model proposed a hierarchical organization:
CSC → self-renewal → differentiated cancer-cell populations → tumor heterogeneity
This model helped explain why some tumors can regenerate after apparently successful treatment. However, research using lineage tracing, single-cell analysis and studies of tumor plasticity has complicated this picture.
In many cancers, differentiated or non-CSC cells can acquire stem-like characteristics. Conversely, CSC-like cells can move toward more differentiated states. This creates a continuum rather than a rigid hierarchy.
2. How Common Are Cancer Stem Cells?
There is no single percentage that describes how common CSCs are across all cancers.
Published studies have reported CSC-like populations ranging from extremely small fractions of tumor cells to populations representing a substantial proportion of the tumor. The reported frequency depends on:
- the cancer type and subtype;
- the patient and tumor stage;
- the CSC marker being used;
- whether cells are measured in primary tumors, metastases or cell lines;
- the experimental assay used;
- whether CSCs are defined by a surface marker or by functional tumor-initiating ability;
- the tumor microenvironment; and
- the biological state of the tumor at the time of sampling.
Cancer stem cells (CSCs) are probably more common than the old “rare cell” model suggested—but there is no single percentage that applies to all cancers.
Published estimates vary enormously, roughly from <1% to >80% of tumor cells, depending on cancer type, sample, definition, biomarkers, and experimental method. A 2024 review reported CSC proportions from 0.2% to 82.5% across primary tumors. DOI
Some examples:
- Acute myeloid leukemia (AML): often <1% using traditional functional definitions.
- B-cell precursor acute lymphoblastic leukemia (ALL): reported CSC-like populations can reach ~82.5%.
- Lung cancer: CD133-positive populations have been reported around 0.4–1.5%, although CD133 alone is not a definitive CSC marker.
- Colorectal cancer: some studies report CSC populations approaching ~20%, depending on the marker and assay.
- Breast cancer: estimates can vary substantially, with higher CSC-like fractions reported in aggressive subtypes. DOI
The important distinction
The statement “CSCs are only 1% of cancer cells” is misleading.
Early transplantation experiments often identified extremely small populations capable of initiating tumors in immunodeficient mice. For example, historical AML experiments estimated frequencies around 0.0004–0.001% under particular experimental conditions. Wiley Online Library
But newer research suggests that stemness is not necessarily a permanent identity restricted to a tiny population.
Cancer cells can undergo phenotypic plasticity:
non-CSC → stem-like/CSC state → differentiated cancer cell
This means a treatment that eliminates the currently detectable CSC population may not eliminate the capacity to regenerate CSC-like cells. Frontiers
This changes how we should think about CSC inhibitors
The more modern model is therefore:
CSC ≠ one rare, fixed cell type
Instead:
CSC = a tumor cell state characterized by properties such as self-renewal, tumor initiation, plasticity and therapy resistance.
That distinction is particularly important for your previous question about CSC inhibitors. A drug that eliminates CD133+, CD44+, ALDH-high or another marker-defined population does not automatically eliminate all cancer stemness.
Indeed, a 2024 review notes that CSC markers such as CD133 have significant limitations and that CSC proportions can vary dramatically between patients and tumor types. DOI
Bottom line: CSCs can be rare, relatively common, or potentially represent a substantial fraction of a tumor, depending on the cancer and how “CSC” is defined. The increasingly important concept for 2026 is CSC plasticity, not simply CSC rarity.
Consequently, claims such as “CSCs make up only 1% of cancer cells” are an oversimplification.
2.1 Why reported CSC percentages vary so much
CSC markers such as CD133, CD44, ALDH and others are useful research tools, but they are not universal identifiers of CSCs. A marker-positive cell does not automatically demonstrate all of the functional properties associated with cancer stemness.
For example, CD133-positive populations can vary substantially between tumors, and CD133 expression by itself does not prove that every positive cell is a tumor-initiating CSC.
Similarly, ALDH-high cells are frequently used as a marker of stem-like populations in several cancers, but ALDH activity also has biological functions unrelated to CSC identity.
2.2 The old “rare CSC” model versus the modern plasticity model
Historically, some transplantation experiments suggested that tumor-initiating cells could be extremely rare. These findings supported the idea of a small CSC compartment at the apex of a cellular hierarchy.
More recent work indicates that this is not the whole story.
Cancer cells can undergo phenotype switching and acquire stem-like properties in response to genetic, epigenetic and environmental signals. This means that even if a treatment eliminates a currently identifiable CSC population, other tumor cells may potentially acquire stem-like properties later.
The therapeutic problem may not simply be “How do we kill the CSCs?” but rather:
“How do we prevent cancer cells from maintaining or reacquiring the stem-like state?”
3. CSC Plasticity: Why It Matters for Cancer Treatment
Cancer-cell plasticity describes the ability of tumor cells to change phenotype in response to intrinsic and environmental signals.
A simplified model is:
Non-CSC ↔ stem-like state ↔ differentiated state
This reversibility is clinically important because chemotherapy, targeted therapy, radiation, hypoxia and other stresses can impose selective pressure on a tumor.
Cells capable of entering a drug-tolerant or stem-like state may survive the initial treatment. Those surviving cells can subsequently repopulate the tumor.
Research has connected CSC plasticity with:
- epithelial-mesenchymal transition (EMT);
- drug-tolerant persister states;
- metastatic dissemination;
- tumor recurrence;
- hypoxia;
- epigenetic reprogramming;
- metabolic adaptation;
- immune evasion; and
- tumor-microenvironment signaling.
Recent reviews describe CSCs and non-CSCs as dynamically interconvertible populations rather than completely separate categories. [1–4]
4. Why Are CSCs Associated With Treatment Resistance?
CSC-like populations may possess several characteristics that allow them to survive conventional treatment.
4.1 Enhanced DNA damage responses
Some stem-like cancer populations have enhanced mechanisms for repairing or tolerating treatment-induced DNA damage.
4.2 Drug efflux
ATP-binding cassette transporters and related mechanisms can reduce intracellular concentrations of some anticancer drugs.
4.3 Quiescence
Some CSC-like cells may divide relatively slowly. This can reduce their sensitivity to therapies that preferentially affect rapidly proliferating cells.
4.4 Metabolic flexibility
CSC populations can display distinctive metabolic programs involving mitochondrial function, glycolysis, oxidative metabolism and redox control.
4.5 Tumor-microenvironment protection
Hypoxia, cancer-associated fibroblasts, extracellular matrix components, cytokines and immune cells can contribute to the maintenance of stem-like phenotypes.
5. Major Cancer Stem Cell Inhibitor Targets
CSC-directed drug development generally focuses on pathways that regulate self-renewal, survival, differentiation and plasticity.
5.1 Wnt/β-catenin
Wnt signaling is one of the most extensively studied pathways in CSC biology. Abnormal Wnt/β-catenin activity can promote self-renewal and stem-like phenotypes in several cancers.
Experimental approaches include:
- porcupine inhibitors;
- β-catenin/CBP pathway inhibitors;
- Frizzled-targeting strategies; and
- other approaches that interfere with Wnt ligand secretion or signaling.
Examples investigated in research include WNT974/LGK974 and PRI-724/ICG-001-related approaches.
Evidence status: primarily investigational for broad CSC eradication. Wnt signaling is biologically compelling, but pathway inhibition must account for pathway redundancy and potential toxicity.
5.2 Hedgehog
Hedgehog signaling regulates developmental and stem-cell programs and has been implicated in CSC maintenance in several malignancies.
Drugs targeting this pathway include:
- vismodegib;
- sonidegib; and
- glasdegib.
These drugs demonstrate an important distinction: a pathway inhibitor can be an established treatment for a particular cancer indication without being a universally proven “CSC killer.” Their clinical use depends on the specific cancer and regulatory indication.
5.3 Notch
Notch signaling participates in cell-fate determination and has been associated with CSC maintenance, differentiation, EMT and therapy resistance.
Investigational approaches include γ-secretase inhibition and antibodies directed against Notch ligands or receptors.
The major challenge is that Notch signaling is also important in normal tissues, making therapeutic selectivity important.
5.4 PI3K/AKT/mTOR
The PI3K/AKT/mTOR network integrates growth signals, nutrient availability and cellular survival. Dysregulation can support cancer-cell survival and stem-like phenotypes.
Agents affecting this pathway include PI3K inhibitors, AKT inhibitors and mTOR inhibitors such as everolimus.
Because this pathway is also central to normal cellular physiology, toxicity and adaptive pathway reactivation remain important challenges.
5.5 JAK/STAT3
STAT3 can connect inflammatory signaling, survival, immune regulation and stemness. Persistent STAT3 activation has been implicated in CSC maintenance in several cancers.
STAT3 and JAK inhibition therefore remains an area of translational research, particularly in tumors with strong inflammatory signaling.
5.6 NF-κB
NF-κB signaling can promote inflammatory survival programs and has been linked to stemness, EMT and resistance to therapy.
However, systemic NF-κB inhibition is complicated by the pathway's essential roles in normal immune function.
5.7 TGF-β
TGF-β is a major regulator of EMT, invasion and cellular plasticity. In some tumor contexts, TGF-β signaling can facilitate acquisition of stem-like characteristics.
Targeting TGF-β therefore represents a strategy aimed not only at CSCs themselves but also at the processes that allow differentiated tumor cells to acquire stem-like properties.
5.8 YAP/TAZ and Hippo signaling
YAP and TAZ integrate mechanical, metabolic and extracellular signals. Abnormal activation has been associated with stemness, regeneration-like programs and resistance to therapy.
YAP/TAZ inhibition is an emerging area rather than an established general CSC treatment.
5.9 ALDH
High aldehyde dehydrogenase activity is frequently used to identify stem-like cancer populations. Experimental ALDH inhibition has therefore attracted interest.
However, ALDH is a family of enzymes with important normal biological functions. Marker-based targeting remains considerably more complicated than simply eliminating every ALDH-high cell.
5.10 CD44 and CD133
CD44 and CD133 are among the most frequently studied CSC-associated surface markers.
Experimental approaches include antibodies, antibody-drug conjugates and cellular therapies.
The major limitation is that these markers are not exclusive to CSCs and may also be expressed by normal or non-CSC populations.
6. Repurposed Drugs and CSC Research
A number of established drugs have been investigated for effects on CSC-related pathways. This is scientifically interesting because repurposed medicines may have known pharmacology and established manufacturing pathways.
 |
| Credit: paulmarik.substack.com |
However, most evidence for the following agents remains preclinical or early translational evidence when the specific claim is CSC elimination.
Metformin
Metformin has been investigated in cancer research because of effects involving cellular energy metabolism, AMPK signaling, insulin signaling and potentially stem-like tumor populations.
Observational studies and laboratory experiments have generated interest, but metformin should not be presented as an established CSC treatment.
Statins
Statins have attracted research interest because cholesterol and mevalonate-pathway biology can influence tumor signaling, membrane organization and potentially stemness.
Clinical evidence remains insufficient to establish statins as general-purpose CSC inhibitors.
Disulfiram
Disulfiram has demonstrated anticancer and CSC-related effects in laboratory models, including interactions involving ALDH and metal-dependent mechanisms.
Clinical translation remains uncertain and should not be confused with an established CSC therapy.
Niclosamide
Niclosamide has attracted considerable oncology research interest because it can influence multiple signaling pathways, including Wnt/β-catenin and other networks associated with cancer-cell survival and stemness.
Its limitations include pharmacokinetic and formulation challenges, and clinical evidence for CSC eradication remains insufficient.
Ivermectin and benzimidazoles
Ivermectin, mebendazole and fenbendazole have been investigated in preclinical cancer models, including studies involving pathways potentially relevant to stemness, proliferation, metabolism and drug resistance.
7. CSC Inhibition Is Probably Not Enough by Itself
One of the most important implications of CSC plasticity is that eliminating a marker-defined CSC population may not permanently eliminate stemness.
A tumor can potentially respond through:
- activation of alternative signaling pathways;
- phenotypic switching;
- EMT or partial EMT;
- metabolic adaptation;
- microenvironmental protection;
- epigenetic reprogramming; and
- selection of resistant clones.
This creates a strong rationale for combination strategies.
The future of CSC therapy may be less about finding one “CSC killer” and more about blocking the tumor's ability to regenerate stem-like, therapy-resistant states.
8. The Emerging Combination-Therapy Model
A more sophisticated treatment framework could involve three interconnected targets:
1. Tumor driver
Target the dominant oncogenic dependency.
2. Resistant/stem-like state
Block pathways supporting stemness, survival or drug tolerance.
3. Tumor microenvironment
Interrupt signals that allow cancer cells to survive, adapt or reacquire stem-like characteristics.
Conceptually:
Oncogenic driver inhibition + CSC/stemness targeting + microenvironment/plasticity control
This approach is particularly relevant to cancers in which acquired resistance is a major clinical problem.
9. CSCs and Targeted-Therapy Resistance
Targeted therapies can produce dramatic initial responses while leaving behind residual cancer cells. Some residual populations can enter drug-tolerant states that subsequently evolve into stable resistance.
CSC biology provides one possible framework for understanding this phenomenon, but CSCs should not be assumed to explain every case of acquired resistance.
Resistance can also arise through:
- secondary mutations;
- bypass signaling;
- gene amplification;
- histologic transformation;
- epigenetic adaptation;
- phenotypic plasticity; and
- selection of pre-existing resistant clones.
Modern precision oncology therefore needs to integrate CSC biology with genomic, transcriptomic and clinical resistance mechanisms.
10. CSCs, Metastasis and Recurrence
CSC-like properties have been associated with invasion, metastatic colonization and recurrence in several experimental systems.
However, it would be inaccurate to say that CSCs are the sole cause of metastasis or recurrence. Cancer progression is an ecosystem-level process involving tumor genetics, immune cells, stromal cells, extracellular matrix, vascular biology and systemic factors.
CSC plasticity may provide one mechanism through which tumor cells adapt to new environments.
11. What Would a Truly Effective CSC Therapy Need to Do?
An effective CSC-directed therapy would ideally demonstrate several properties:
- selective targeting of tumor stem-like cells;
- minimal damage to normal stem/progenitor cells;
- prevention of CSC regeneration from non-CSC populations;
- suppression of compensatory pathway activation;
- activity against resistant and metastatic disease;
- demonstrable pharmacodynamic effects in humans; and
- improved clinical outcomes such as progression-free or overall survival.
Most proposed CSC inhibitors have not yet satisfied all of these criteria.
12. Evidence Hierarchy for CSC Inhibitors
| Evidence level | What it means | Interpretation |
|---|---|---|
| Level 1 | Laboratory mechanism | Drug affects a CSC-associated pathway or marker in vitro. |
| Level 2 | Animal evidence | CSC-related effects demonstrated in animal models. |
| Level 3 | Early human studies | Safety, pharmacology or biomarker effects studied in patients. |
| Level 4 | Clinical efficacy evidence | Controlled studies demonstrate meaningful clinical benefit. |
| Level 5 | Validated clinical therapy | Evidence supports routine use for a defined cancer indication. |
Many compounds described online as “CSC inhibitors” remain at Levels 1–2. This distinction is essential when evaluating repurposed drugs and natural compounds.
13. Natural Compounds and CSC Research
Several dietary compounds have been investigated experimentally for effects on pathways associated with cancer stemness.
Examples include:
- curcumin;
- epigallocatechin gallate (EGCG);
- sulforaphane;
- resveratrol; and
- quercetin.
These compounds can influence signaling pathways in laboratory systems, but laboratory concentrations, bioavailability and human pharmacology can differ substantially from real-world dietary exposure.
Therefore, evidence that a natural compound inhibits a CSC pathway in a cell culture experiment should not be interpreted as evidence that taking a supplement will eliminate CSCs in a patient.
14. The Biggest Problem: CSC Heterogeneity
There may not be one universal CSC phenotype.
Different cancers—and even different regions of the same tumor—can contain different stem-like states.
CSC biology can also change over time following:
- chemotherapy;
- radiation;
- targeted therapy;
- immunotherapy;
- hypoxia;
- nutritional changes; and
- changes in the tumor microenvironment.
This means that a future CSC treatment may need to be dynamic and biomarker-guided rather than based on one universal CSC marker.
15. What Is the Most Promising Direction?
The strongest conceptual direction is moving from the idea of a single CSC population toward a model of tumor-cell state transitions.
Under this model, effective therapy may require simultaneous control of:
- oncogenic signaling;
- stemness pathways;
- cellular plasticity;
- metabolic adaptation;
- epigenetic state;
- immune evasion; and
- the tumor microenvironment.
Single-cell sequencing, spatial transcriptomics, lineage tracing and longitudinal liquid-biopsy approaches may eventually make it possible to identify which resistant cellular states emerge during treatment.
16. Frequently Asked Questions
Are cancer stem cells real?
Yes. CSC-like populations with tumor-initiating and self-renewal properties have been demonstrated in multiple cancers. However, the definition, frequency and biological importance of CSCs vary between tumor types and experimental systems.
Are cancer stem cells only 1% of a tumor?
No. There is no universal 1% rule. Reported frequencies vary substantially depending on the cancer, marker and experimental methodology. More importantly, non-CSC cells may acquire stem-like properties through cellular plasticity.
Can cancer stem cells be eliminated?
Researchers are actively investigating this, but there is currently no universal therapy proven to eradicate CSCs across all cancers.
What is the best cancer stem cell inhibitor?
There is no single best CSC inhibitor. Candidate therapies depend on the cancer type and the dominant biological pathway. Wnt/β-catenin, Hedgehog, Notch, PI3K/AKT/mTOR, JAK/STAT3, TGF-β and YAP/TAZ are among the major areas of research.
Does metformin kill cancer stem cells?
Metformin has demonstrated CSC-related effects in laboratory and translational research, but it has not been established as a general CSC-eradicating cancer treatment.
Does ivermectin kill cancer stem cells?
Some laboratory studies have reported effects of ivermectin on cancer-related signaling and cellular processes. However, there is currently insufficient clinical evidence to establish ivermectin as a CSC treatment or standard cancer therapy.
Does fenbendazole kill cancer stem cells?
Fenbendazole has generated interest because of laboratory anticancer findings, but evidence that it selectively eliminates CSCs in human cancer patients is not established. It should be regarded as investigational in oncology.
Why might CSC inhibitors need to be combined with standard treatment?
Because CSC-like cells represent only one component of tumor biology. Effective cancer treatment may require simultaneous control of the dominant tumor driver, bulk tumor population, resistant states and microenvironment.
17. Conclusion
Cancer stem cells represent an important—and increasingly sophisticated—area of cancer research. They may contribute to tumor initiation, heterogeneity, metastasis, recurrence and treatment resistance, but the simple concept of a tiny, permanently fixed CSC population is no longer sufficient.
One of the most important developments is the recognition of CSC plasticity. Cancer cells can potentially transition between stem-like and differentiated states, meaning that eliminating a currently identifiable CSC population may not be sufficient to prevent tumor regeneration.
This changes the therapeutic objective.
- targeting stemness pathways;
- preventing re-acquisition of stem-like states;
- blocking adaptive resistance;
- targeting the tumor microenvironment;
- combining CSC-directed strategies with established cancer therapy; and
- using biomarkers and longitudinal monitoring to identify evolving resistant states.
For patients, the distinction between preclinical promise and proven clinical benefit remains critical. A drug that inhibits a CSC-associated pathway in laboratory experiments is not automatically an effective CSC treatment in humans.
The most scientifically defensible position in 2026 is therefore not that one drug can “kill all cancer stem cells,” but that stemness, plasticity and therapy-resistant cellular states represent potentially actionable components of cancer biology that may eventually be targeted through personalized combination therapy.
References
- Bhat GR, Sethi I, Sadida HQ, et al. Cancer cell plasticity: from cellular, molecular, and genetic mechanisms to tumor heterogeneity and drug resistance. Cancer Metastasis Reviews. 2024;43:197–228.
- Knopik-Skrocka A, Sempowicz A, Piwocka O. Plasticity and resistance of cancer stem cells as a challenge for innovative anticancer therapies – do we know enough to overcome this? EXCLI Journal. 2024;23:335–355.
- Higa T, Nakayama KI. Cell cycle heterogeneity and plasticity of colorectal cancer stem cells. Cancer Science. 2024;115:1370–1377.
- Cancer stem cell hypothesis 2.0 in glioblastoma: Where are we now and where are we going? Neuro-Oncology. 2024;26:785–797.
- Interplay between epigenetics and metabolism controls cancer stem cell plasticity. Frontiers in Epigenetics and Epigenomics. 2024.
- Targeting cancer stem cell plasticity and tumor microenvironment crosstalk: a comprehensive review. 2025/2026 review literature.
- Targeting the Mitochondrial–Stem Cell Connection in Cancer. Journal of Orthomolecular Medicine. 2024.
- The Evolution of Cancer Resistance: How Tumors Outsmart Therapy and Practical Strategies to Stay One Step Ahead. Dr Paul Marik, MD. 2026.
Medical Disclaimer
This article is for educational and research purposes only and is not medical advice. Cancer stem cell inhibitors discussed here include experimental and investigational approaches. Preclinical findings do not establish clinical efficacy. Patients should not replace chemotherapy, targeted therapy, immunotherapy, surgery, radiation or other evidence-based cancer treatment with experimental or repurposed drugs without discussing the decision with a qualified oncology team.
In particular, ivermectin, fenbendazole, mebendazole, niclosamide, metformin, disulfiram and natural compounds should not be interpreted as proven cancer stem cell treatments solely because laboratory studies demonstrate effects on CSC-associated pathways.
Editorial note: This article should be updated as clinical trials establish whether CSC-directed or plasticity-directed strategies improve patient outcomes. The distinction between pathway inhibition, CSC-marker reduction and actual clinical eradication of therapy-resistant disease should be maintained in all future updates.
.png)
.png)
Comments
Post a Comment