Stem Cells & Cancer

Note: This page is educational and reflects the state of the literature in 2025. It does not replace medical advice.

TL;DR

Most tumors are not uniform — they contain a small subpopulation of cells with stem-like properties (self-renewal, differentiation, tumor-initiation) called cancer stem cells (CSCs). CSCs are believed to drive tumor initiation, relapse, metastasis, and therapy resistance. Understanding them reframes "killing the tumor" as "killing the cells that re-grow the tumor."

1. From normal stem cells to cancer stem cells

A normal stem cell has two defining properties:

Self-renewal — divides to produce another stem cell.
Differentiation — divides to produce specialized progeny (e.g., a hematopoietic stem cell yields red and white blood cells).

These same properties, hijacked by mutation and microenvironmental cues, are the working definition of cancer stem cells (CSCs) — also called tumor-initiating cells. CSCs sit at the apex of an intratumoral hierarchy: a small fraction of cells re-seeds the tumor; the bulk are more differentiated and have limited proliferative capacity. Sources: [1], [2]

Skeptic's corner. "CSC" is a functional definition (cells that initiate a tumor when transplanted at low number into an immunocompromised host). Marker-based identification (CD44, CD133, ALDH activity, etc.) is convenient but imperfect — markers vary by tumor type and even within a tumor. Plasticity blurs the line further. Sources: [2]

2. Why the CSC model matters clinically

Three predictions of the CSC model have practical consequences:

Resistance to therapy. Standard chemotherapy and radiotherapy hit dividing cells hardest. Quiescent CSCs survive and re-seed the tumor. This is consistent with relapse patterns in glioblastoma, AML, and breast cancer. Sources: [1], [3]
Metastasis. Disseminated tumor cells with CSC features are enriched at metastatic sites; CSC pathways (Wnt, Notch, Hedgehog, EMT-associated transcription factors) overlap heavily with metastasis programs. Sources: [1]
Relapse. Even tumors that respond impressively to first-line therapy can return — the surviving fraction is enriched for stem-like properties. Sources: [1]

Practical implication: a therapy that shrinks a tumor on imaging but spares CSCs may still fail. This is why combinations targeting both the bulk and the CSC compartment are an active field. Sources: [1]

3. The signaling pathways most often cited

CSC self-renewal is linked to a relatively small set of pathways, repeatedly observed across tumor types: Sources: [1]

Wnt/β-catenin — colorectal, breast, leukemia.
Notch — T-cell ALL, glioblastoma, breast.
Hedgehog (SHH) — basal cell carcinoma, medulloblastoma, pancreas.
NF-κB / JAK-STAT / TGF-β / PI3K-AKT — broad roles in stemness and the inflammatory niche.
Epigenetic regulators (BMI1, EZH2, DNMTs) — maintain the chromatin state that allows self-renewal.

These pathways are also normal developmental pathways — drugging them is a balancing act between hitting CSCs and disturbing normal stem cell maintenance (gut crypt, bone marrow, hair follicle).

4. Glioblastoma as a case study

Glioblastoma stem cells (GSCs) are the textbook example — they showed that the CSC concept extends beyond hematological malignancies to solid tumors.GSCs: Sources: [3]

Reside in perivascular and hypoxic niches inside the tumor.
Resist temozolomide and radiation more than the bulk.
Drive systemic immunosuppression that limits checkpoint-inhibitor responses.
Exhibit transcriptional plasticity — single-cell RNA-seq shows GSCs sliding between developmental states rather than sitting in a single "stem" identity. Sources: [3]

This last point is important: CSC identity is dynamic, not fixed. A non-stem cancer cell can re-acquire stemness under stress (therapy, hypoxia, microenvironmental signals) — de-differentiation. This challenges any therapy that aims to eradicate a small, fixed CSC population. Sources: [2]

5. How CSCs are studied (the technical bench-to-bytes view)

For the technologist reader, CSC research is a particularly data-heavy subfield:

Functional assays — limiting dilution transplants in immunocompromised mice; tumorsphere formation in serum-free non-adherent culture; lineage tracing in genetically engineered models. Sources: [2]
Single-cell RNA-seq — to dissect intratumoral hierarchies and detect rare stem-like populations. Tools: Seurat, Scanpy, Slingshot/Palantir for trajectory inference. Background: see Single-cell & spatial intro.
Spatial transcriptomics — to anchor CSC niches (perivascular, hypoxic) anatomically.
CRISPR screens — to identify CSC-specific dependencies (e.g., Project DepMap data).
Drug response surfaces — synergy/antagonism between CSC-targeting agents and standard cytotoxics.

If you are a bioinformatician new to this, the entry path is Seurat → trajectory inference → integrate with spatial. The hardest part is the biology, not the code: reasoning about what "stemness score" actually means in a noisy tumor scRNA-seq dataset.

6. Therapeutic strategies targeting CSCs (research-stage)

Approaches discussed in the literature, with caveats: Sources: [1]

Pathway inhibitors — Wnt (porcupine inhibitors), Notch (γ-secretase inhibitors), Hedgehog (vismodegib, sonidegib — approved for basal cell carcinoma).
Differentiation therapy — force CSCs to differentiate and lose self-renewal (paradigm: ATRA in acute promyelocytic leukemia; harder to reproduce in solid tumors).
Niche disruption — anti-angiogenics (bevacizumab) target the perivascular niche; hypoxia-activated prodrugs target hypoxic niches.
Immune targeting — CAR-T or bispecifics against CSC-enriched antigens (e.g., CD19 in B-ALL works because nearly all leukemic clones express CD19; matching this in solid tumors is harder).
Metabolic targeting — CSCs often have distinct mitochondrial/ROS dependencies vs. bulk tumor.

Most CSC-targeting strategies are still research-stage for solid tumors. Approved CSC-related therapies (vismodegib, ATRA, gemtuzumab) are exceptions, not the rule. Sources: [1]

7. Open questions

Is "CSC" a state cells transition into (plasticity) or a fixed sub-population? In most solid tumors, evidence points to state, not lineage. Sources: [2]
How much of "stemness" is intrinsic vs. instructed by the niche?
Can we image CSCs in patients (PET tracers, liquid biopsy)?
Do conventional response criteria (RECIST) miss CSC-relevant outcomes?

References

Chu X, Tian W, Ning J, et al. Cancer stem cells: advances in knowledge and implications for cancer therapy. Signal Transduct Target Ther 2024;9:170. PMID 38965243. https://doi.org/10.1038/s41392-024-01851-y
Nassar D, Blanpain C. Cancer Stem Cells: Basic Concepts and Therapeutic Implications. Annu Rev Pathol 2016;11:47-76. PMID 27193450. https://doi.org/10.1146/annurev-pathol-012615-044438
Gimple RC, Bhargava S, Dixit D, Rich JN. Glioblastoma stem cells: lessons from the tumor hierarchy in a lethal cancer. Genes Dev 2019;33:591-609. PMID 31160393. https://doi.org/10.1101/gad.324301.119
National Cancer Institute (NCI). What is cancer? https://www.cancer.gov/about-cancer/understanding/what-is-cancer
American Cancer Society. Cancer A-Z (overview). https://www.cancer.org/cancer.html
Cleveland Clinic. Cancer (overview). https://my.clevelandclinic.org/health/diseases/12194-cancer
A.C. Camargo Cancer Center. O que é câncer. https://accamargo.org.br
Fundação do Câncer (Brasil). https://www.cancer.org.br/

Stem Cells & Cancer ​

TL;DR ​

1. From normal stem cells to cancer stem cells ​

2. Why the CSC model matters clinically ​

3. The signaling pathways most often cited ​

4. Glioblastoma as a case study ​

5. How CSCs are studied (the technical bench-to-bytes view) ​

6. Therapeutic strategies targeting CSCs (research-stage) ​

7. Open questions ​

See also ​

References ​