Abstract
Emerging evidence suggests that cancer stem cells account for the initiation and progression of cancer. While many types of cancer stem cells with specific markers have been isolated and identified, a variety of differences among them began to be appreciated. Cancer stem cells are hierarchical populations that consist of precancerous stem cells, primary cancer stem cells, migrating cancer stem cells and chemoradioresistant cancer stem cells, playing different roles in cancer initiation and progression. Here we propose a new concept "horizontal hierarchy of cancer stem cells" to distinguish them from vertical hierarchy cancer stem cells, cancer transient-amplifying cells and cancer differentiated cells, and summarize our current understanding of these subsets of cancer stem cells with the aim to open up novel therapeutic strategies for cancer based on this understanding.
Introduction: Beyond the Static CSC Model
The "Cancer Stem Cell Hypothesis" posits that a small subpopulation of cells drives tumor growth and recurrence. Traditional oncology treatments often fail because they target rapidly dividing "bulky" tumor cells while sparing these resilient, often quiescent, cancer stem cells (CSCs). However, a significant challenge in the field is the tendency to treat all CSCs as a monolithic population.
This paper addresses this oversimplification by highlighting that CSCs are in a state of constant evolution. A CSC that initiates a primary tumor is biologically distinct from one that undergoes epithelial-mesenchymal transition (EMT) to seed distant organs, or one that survives intensive chemoradiotherapy. Understanding these functional differences is critical for developing therapies that can truly eradicate the disease.
The Horizontal Hierarchy Model
To better describe the evolution within the stem cell compartment, the authors propose a "horizontal hierarchy" framework. While the traditional "vertical" hierarchy describes the differentiation path from a CSC to a non-tumorigenic differentiated cell, the horizontal model maps the progression of the stem cells themselves through various functional states.
As shown in Figure 3, the model tracks the transition from normal stem cells (SCs) to precancerous stem cells (pre-CSCs), primary cancer stem cells (pri-CSCs), and eventually to specialized subsets like migrating (mig-CSCs) and resistant (cr-CSCs/rr-CSCs) populations. This progression mirrors the clinical stages of cancer from precancerous lesions to metastasis and relapse.
Functional Subsets and Markers
The identification of these subsets relies heavily on specific cell surface markers and signaling pathways. Through an extensive review of over 100 studies, the authors synthesized a comprehensive list of markers used to isolate CSCs across nearly 20 different cancer types.
Key insights from this meta-analysis include:
- Pre-CSCs: Often dormant within benign lesions, these cells express "stemness" factors like OCT3/4 and SOX2 but require micro-environmental triggers (e.g., hypoxia) to progress.
- Migrating CSCs: These cells are typically the product of EMT, characterized by a loss of E-cadherin and a gain of Vimentin. In pancreatic cancer, for instance, the CD133+/CXCR4+ phenotype is specifically linked to this metastatic subset.
- Resistant CSCs: These subsets survive therapy through high expression of ABC transporters (drug efflux pumps), lower levels of reactive oxygen species (ROS), and enhanced DNA repair mechanisms.
Clinical Implications and the Relapse Loop
The horizontal hierarchy explains why traditional therapies often lead to relapse. While drugs may successfully eliminate the "vertical" descendants (differentiated tumor cells) and even some primary CSCs, the specialized migrating and resistant subsets often remain. These surviving cells can then re-initiate tumor growth or seed new metastases, creating a cycle of recurrence. Effective treatment must therefore target the specific vulnerabilities of these horizontal subsets or the transitions between them.
Future Directions and Limitations
While the horizontal hierarchy provides a robust conceptual framework, several challenges remain. Many of the proposed transitions between subsets are currently speculative and require direct experimental validation through lineage tracing in human models. Furthermore, the overlap of markers between different subsets makes clinical isolation difficult.
Future research should focus on identifying unique molecular signatures for each functional state and developing therapeutic agents that can specifically target the "plasticity" of CSCs—preventing, for example, the transition of a primary CSC into a migrating or resistant one. This represents a major frontier in the quest for curative cancer treatments.
Summary updated