Cancer stem cell mechanisms and targeted therapeutic strategies in head and neck squamous cell carcinoma

Abstract

Head and neck squamous cell carcinoma (HNSCC) originates in the epithelial lining of the oral cavity, pharynx, and larynx, with over 830,000 new cases diagnosed globally in 2020, making it the seventh most prevalent cancer. Despite treatment advances, high-grade HNSCCs remain associated with poor outcomes and a high risk of recurrence.

Although Cancer Stem Cells (CSCs) are rare in HNSCC tumors, they are key drivers of tumor relapses, as they evade apoptosis and survive current therapies through enhanced DNA repair and quiescence. This review integrates recent discoveries into a unified framework for understanding CSC mechanisms. It examines the role of pluripotency factors, biomarkers, replicative immortalization, metabolic reprogramming, redox regulation, and immune evasion in shaping CSC behavior and survival under treatment stress. Non-coding RNAs are also discussed as modulators of gene expression via epigenetic regulation in CSCs. Importantly, it highlights how these adaptive mechanisms intersect as potential vulnerabilities that could be exploited to eliminate CSCs through targeting multiple pathways.

Finally, it provides recent and emerging precision therapeutics, including CAR-T cells, immune checkpoint inhibitors, metabolic and redox-targeting agents, and epigenetic modulators currently in preclinical and clinical trials applications designed to eliminate CSCs and improve patient outcomes in HNSCC.

1. Introduction

Head and neck squamous cell carcinoma (HNSCC) arises from the epithelial tissues of the oral cavity, pharynx, or larynx (Fig. 1A). Treatment typically begins with surgical removal of the tumor (Fig. 1B), followed by biopsy and pathological evaluation (Fig. 1C), and staging based on the TNM system (tumor size, lymph node, and metastasis) [1]. These results guide radiation and chemotherapy (Fig. 1D). While early-stage HNSCC outcomes have improved, up to 50 % of advanced HNSCC recur or metastasize despite therapy [2,3]. Research has illustrated the intricate mechanisms governing the behavior and treatment resistance of cancer stem cells (CSCs) in HNSCC, which can generate new tumors and lead to cancer relapse (Fig. 1E).

Fig. 1.

Current treatment of primary HNSCC tumor involves a multimodal approach.

A, B. HNSCC tumors are surgically removed when feasible, followed by C. pathological assessment to determine tumor margins, lymph node involvement, and staging. A team of surgical oncologists, radiation oncologists, and medical oncologists then tailors the next steps. D. Depending on factors such as tumor location, stage, and margin status, adjuvant therapy may include radiation therapy alone or combined with chemotherapy, particularly in cases with high-risk features such as positive margins, extracapsular extension, or perineural invasion. E. Cancer stem cells can remain quiescent, resist treatments, and with time can generate new and more aggressive tumors.

2. Models for understanding cell heterogeneity in HNSCC tumors

Like in other cancers, HNSCC tumor cells invade surrounding tissues and form tumor nests with diverse cancer cell types, making treatment difficult due to the lack of universal targets that spare healthy tissue. Several models have been proposed to explain tumor heterogeneity. In the normal stem cell model, a single cell can self-renew and divide into identical stem cells or give rise to diverse differentiated cell types (Fig. 2A). The CSC/Hierarchical model suggests that CSCs can divide or differentiate, generating a heterogeneous mix of cells, contributing to tumor complexity and diversity [4–6] (Fig. 2B). The clonal evolution model, or stochastic model, proposes that cancer cells accumulate mutations leading to clonal diversity as they replicate [6]. Like in evolution, variants are continually selected based on their replication capacity and survival fitness [7]. This results in multiple sub-clones that drive tumor progression and heterogeneity (Fig. 2C). Finally, the hybrid CSC model (or plasticity model) replaced the rigid hierarchy with a dynamic system. Here, tumor cells can acquire stem-like traits through epigenetic reprogramming and dedifferentiate in response to microenvironmental signals, reactivating stemness pathways [8] (Fig. 2D). In all models, disrupting stemness mechanisms offers a potential strategy for intervention.

Fig. 2. Models of Normal and Cancer Stem Cell Differentiation and Tumor Evolution.

An illustration of four models of stem cell differentiation and cancer development: A. The Normal Adult Stem Cell Model: Stem cells follow a unidirectional hierarchy, differentiating into progenitor cells, then into fully differentiated cells. B. The Cancer Stem Cell (CSC)/Hierarchical Model: CSCs are rare cells capable of initiating and sustaining tumor growth. CSCs divide asymmetrically, into non-CSCs that cannot revert to a stem-like state. C. The Clonal Evolution/Stochastic Model: All tumor cells undergo random mutations, epigenetic changes, and environmental influences. This leads to heterogeneity, with aggressive and therapy-resistant clones eventually dominating. D. The CSC Plasticity/Hybrid Model: Integrated features of both hierarchical and stochastic theories: Cancer cells can switch between states. Cancer cells can dedifferentiate or acquire stemness properties, contributing to tumor adaptability.

3. Stemness unleashed: pluripotency factors in HNSCC CSCs

Pluripotency is the ability of stem cells to divide and form any tissue type or organ. During cancer cell transformation, embryonic pluripotency factors are reactivated, leading to enhanced plasticity, loss of cell function, and abnormal or partial cell identity. Pluripotency factors such as POU Class 5 Homeobox 1 (POU5F1, aka OCT3/4), SRY-box transcription factor 2 (SOX2), and homeobox protein NANOG, which normally maintain stemness in embryonic stem cells, are active in CSCs of HNSCC (Fig. 3) [9]. These transcriptional factors significantly contribute to the properties of CSCs, including their capacity for self-renewal, dedifferentiation, and resistance to conventional therapies [9]. For example, overexpression of SOX2 has been associated with an increase in CSC frequency, enhanced tumorigenicity, and a poor prognosis [10]. POU5F1 is involved in ABC transporters and anti-apoptotic genes upregulation [11] (Fig. 3), further complicating efforts to eradicate these cells. Moreover, NANOG and POU5F1 promote the invasiveness and metastatic potential of HNSCCs [12,13].

Fig. 3. CSC markers in HNSCC.

A. The classical CSC surface markers include CD133, CD44, and the transporter ABCG2. They are also characterized by high ALDH1A1 expression and the pluripotency factors POU5F1, SOX2, and NANOG. B. 5 % of the population transiently express the early embryonic gene ZSCAN4. This factor induces chromatin remodeling, telomere extension, and an increase in pluripotency-gene expression.

3.1. Markers of CSCs in HNSCC

Unique surface proteins of CSCs are important as they can serve as entry points to deliver drugs. By being more specific, pathologists can also utilize them as markers to identify hidden CSC populations and assist surgeons in more accurately defining tumor margins during surgery.

Well-established CSC extracellular markers of HNSCC are shown in Table 1. For example, cells that express high CD44 are linked to tumor initiation, poor prognosis, and high metastasis rates. Under normal conditions, CD44 mediates adhesion and migration by binding hyaluronic acid on endothelial cells in lymphoid tissues. CSCs hijack this mechanism to enter lymph nodes and facilitate cancer spread [14,15]. Markers used to identify and isolate CSCs are CD10 [16], CD133 (PROM1) [17,18], and low expression of CD24 [19]. Other important membrane proteins, such as Integrin subunit beta 1 (ITGB1), are implicated in treatment resistance and metastasis [20]. Epithelial Cell Adhesion Molecule (EpCAM) is another membrane protein used with other markers to isolate CSCs and maintain the epithelial phenotype (Table 1).

Table 1.

Representative membrane and surface-associated markers used to identify CSCs in HNSCC.

Marker	Location	CSC Enrichment Potential (Rated High to Low)	Functional Role	Clinical Relevance
CD44 [21]	Basolateral Membrane	Moderate. (Not exclusive to CSC)	Cell adhesion, migration, and CSC maintenance	Prognostic marker; therapeutic target (anti-CD44 therapies in clinical trials)
CD24 [22]	Apical Membrane	Poor. Discriminatory power	Cell adhesion and tumor progression	Limited prognostic value; coexpressed in various tumors
CD10 [16]	Apical Membrane	Emerging (low specificity)	Metalloproteinase regulation, invasion	Associated with invasion in HNSCC
CD133 [17, 18]	Apical Membrane	High	Stemness and resistance to therapy	Prognostic relevance debated
EpCAM [23]	Cell Membrane	High (CSC enrichment varies)	Epithelial marker, tumor initiation	Marker for epithelial cells
Integrin α6 [24]	Basolateral Membrane	High	Cell adhesion, signaling, and ECM interaction	Associated with self-renewal and tumor initiation
Integrin β1 [25]	Basolateral Membrane	High (variable expression)	Cell adhesion, signaling, and ECM interaction	Regulates stemness via Notch1 signaling; associated with poor prognosis

The table includes markers’ subcellular location, CSC enrichment potential, functional role, clinical relevance, and supporting citations. CSC enrichment potential refers to the marker effectiveness when used in functional assays such as FACS-based sorting, tumor sphere formation, or limiting dilution transplantation in mice. CSC enrichment potential was rated as” High: Consistently detects cells with CSC traits. Moderate: Evidence of enrichment is limited. Low: Limited or inconsistent enrichment across studies in HNSCCs.

Other CSC markers include Aldehyde dehydrogenase 1A1 (ALDH1A1) is a cytosolic enzyme that contributes to the detoxification of chemotherapeutic agents. In HNSCC, high ALDH1A1 activity enriches CSCs with elevated self-renewal, quiescence, and resistance. C-X-C chemokine receptor type 4 (CXCR4) is also implicated in tumor initiation and involved in cell signaling and migration. It binds to CXCL12 to promote metastasis and can be used in combination with CD44 to identify CSC subpopulations [26]. A leucine-rich repeat-containing G-protein-coupled receptor (LGR5) is another CSC factor. LGR5-positive cells have high tumor initiation potential and are associated with treatment resistance [26,27]. Together, these markers enhance the precision of CSC detection and may serve as targets for directed therapies.

3.2. The prognostic value of CSC biomarkers in HNSCC

The clinical relevance of key CSC markers, such as CD133, CD44, and CD24, remains a subject of ongoing debate in HNSCC. These markers can enrich for CSCs in vitro [17] and are associated with tumor aggressiveness and poor outcomes [28]. However, when CD133 is used as a standalone marker, it produces conflicting evidence. Meta-analyses indicate that its prognostic value is either neutral or context-dependent, which limits its clinical application. [29]. Similarly, a meta-analysis shows that while CD44 correlates with poor outcomes in pharyngeal and laryngeal cancers, it is not associated with oral cancer. [30]. To date, most clinical studies of biomarkers are retrospective; no completed randomized trials have been reported that stratify HNSCC patients based on a single CSC surface marker profile. One ongoing multinational non-randomized clinical trial uses soluble CD44 in saliva as a non-invasive prognostic tool to assess the risk of recurrence in oral and oropharyngeal cancer [31]. The prognostic value of CSC markers is higher when evaluated in combination with other markers and using standardized methods [32–34].

4. Sustaining cancer stemness

4.1. Key transcription factors regulating the CSC phenotype in HNSCC

Several transcription factors associated with aggressive tumor behavior are upregulated in HNSCC-CSC. Zinc finger E-box-binding homeobox 1 (ZEB1), a transcription factor involved in epithelial to mesenchymal transition (EMT), was found to be highly expressed in Prominin 1 (PROM1, also known as CD133)-positive cells and critical for the CSC phenotype. [35]. Co-expression of ZEB1 and ZEB2 enhances EMT and stemness via transcriptional repression and predicts poor prognosis [36].

B lymphoma Mo-MLV insertion region 1 homolog (BMI1) [21], is another important transcription factor in CSC and a member of the polycomb group of proteins. BMI1 regulates genes involved in self-renewal of normal stem cells, represses p16 and p19 to maintain CSC properties [21], correlates with poor prognosis and treatment resistance [37].

Other transcription factors acting as EMT drivers are Zinc finger protein (SNAIL) and Twist Family bHLH Transcription Factor (TWIST), which are also highly enriched in CSCs [13]. Notably, SNAIL binds epithelial differentiation genes (E-cadherin/CDH1) and activates mesenchymal programs [38,39]. SNAIL overexpression is linked to tumor recurrence and disease progression in HNSCC [40]. Meanwhile, the concurrent activation of TWIST increases mesenchymal genes (N-cadherin and vimentin) and further represses epithelial traits. Together, they enhance tumor aggressiveness and metastatic spread [41]. This EMT-driven stem-like phenotype is a defining feature of CSCs in HNSCC.

4.2. Immortalization and telomere maintenance mechanisms in HNSCC CSCs

Telomeres are repetitive sequences at the chromosome ends that shorten with each cell division of somatic stem cells and eventually trigger apoptotic cell death. CSCs overcome this barrier by activating telomerase (TERT) to extend the telomeres [42,43]. In HNSCC, mutations in the TERT promoter are enriched in CSCs, supporting replicative immortalization and aggressive tumor behavior [44,45]. While alternative lengthening of telomeres (ALT), a recombination-based mechanism, is rare in HNSCC, it may play a role in certain subsets of HNSCC cells [46].

Zinc finger and SCAN domain containing 4 (ZSCAN4) [47], is a unique stemness factor as it is mostly silent in HNSCC cells (Fig. 3A). When it is transiently expressed, it leads to a robust increase in telomere length [48]. It also increases pluripotency genes and promotes an open chromatin state via DNA demethylation and histone acetylation [48,49] (Fig. 3B). This suggests it drives a transient but potent reprogramming state, functionally enhancing telomere maintenance, epigenetic plasticity, and pluripotency. As embryonic and pluripotency factors are not expressed in normal adult cells, they may serve as excellent targets for cancer therapies.

4.3. Catalyst spread: CSC enzymes driving HNSCC invasion and metastasis

Unlike bulk tumor cells, CSCs can leverage EMT through downregulation of E-cadherin and upregulation of vimentin [13]. This ability increases their motility, invasiveness, and adaptability to microenvironmental stressors. Several factors upregulated in CSCs are involved with extracellular matrix remodeling and cell motility, associated with aggressive tumor behavior and poor prognosis. Notably, matrix metalloproteinases (MMPs), such as Collagenase IV (MMP-2), cleave basement membrane components to facilitate invasion. Accordingly, MMP-2 and MMP-9 expressions correlate with enhanced metastasis and reduced patient survival [50].

4.4. Signaling pathways supporting CSC maintenance and tumor relapse

CSCs harness many pathways to regulate key transcriptional programs that control proliferation, survival under stress, and therapy resistance. Aberrant activation of these pathways reinforces the stemness state.

4.4.1. Sonic Hedgehog (SHH) signaling

Sonic Hedgehog (SHH) signaling is one of the important pathways that maintain CSC properties. Patched (PTCH) is a receptor that inhibits the membrane receptor Smoothened (SMO), preventing its downstream signaling. This allows Protein Kinase A (PKA) and Suppressor of Fused (SUFU) to bind, phosphorylate, and inhibit GLI2/3, preventing their activation (Fig. 4A). When SHH binds to PTCH, it relieves SMO inhibition, triggering downstream activation of GLI transcription factors (GLI1 and GLI2). GLI1 acts as a potent transcriptional activator, while GLI2 regulates activation or repression of key target genes (Fig. 4B). In HNSCC CSCs, SHH-driven GLI transactivates the promoters of NANOG, SOX2, and POU5F1 [10], reinforcing their role in tumor progression and recurrence.

Fig. 4. Hedgehog (HH) signaling pathway and GLI transcription activation of stemness genes.

A. In the absence of Sonic Hedgehog (SHH), Patched (PTCH) inhibits Smoothened (SMO), preventing downstream signaling. Protein Kinase A (PKA) and Suppressor of Fused (SUFU) phosphorylate and inhibit GLI2/3, preventing their activation. B. When SHH binds to PTCH, it relieves SMO inhibition, allowing it to activate downstream signaling. SUFU is inactivated, leading to the activation of GLI2/3, which translocate to the nucleus and initiate transcription of GLI1. Then, GLI1 activates the target genes Cyclin D1/D2, MYC, and BCL2, to increase proliferation and survival, and the factors NANOG, SOX2, and OCT4/POU5F1.

4.4.2. CTGF signaling pathway

Connective Tissue Growth Factor (CTGF) is a multifunctional secreted protein that can function as either an oncoprotein or a tumor suppressor in various cancers [51]. CTGF induces c-Jun expression through the αvβ3 integrin pathway, and c-Jun directly activates the transcription of pluripotency genes, including NANOG, SOX2, and POU5F1 [51]. CTGF and pluripotency gene co-expression are associated with a worse prognosis. FAM3C (FAM3 Metabolism Regulating Signaling Molecule C) is another secreted factor involved in EMT, which enhances cancer-cell plasticity and invasiveness [52].

4.4.3. . The JAK/STAT pathway

The Janus kinase (JAK) activates STAT3 to upregulate PD-L1 and IL-10, which leads to enhanced tumor progression, increases treatment resistance, and promotes immune suppression, shielding CSCs from T-cell-mediated destruction [53].

4.4.4. The Wnt/β-catenin pathway

The Wnt/β-catenin Pathway regulates CSC plasticity. In CSCs, Wnt ligands bind to Frizzled receptors, leading to stabilization and nuclear translocation of β-catenin, which then activates genes involved in stemness, EMT, and cell survival. Elevated β-catenin activity enhances CSC migration and contributes to therapy resistance [54].

4.4.5. The Notch Signaling Pathway

The Notch Signaling Pathway is critical for CSC self-renewal and differentiation. Notch receptors (Notch1–4) are activated through cell-to-cell interactions. This interaction releases Notch intracellular domain (NICD), which then translocates into the nucleus and regulates genes involved in stem cell maintenance, proliferation, and therapy resistance. Notch1 overexpression is associated with increased radiation resistance and poor prognosis in HNSCC [55]. Together, these pathways regulate tumor progression, therapeutic resistance, and metastatic capabilities in HNSCC, positioning them as areas of interest in CSC-directed therapies.

4.5. CSC in human papillomavirus (HPV) positive and negative HNSCC

Human papillomavirus (HPV), particularly HPV16, is a significant cause of oropharyngeal cancers, by expressing the viral oncoproteins E6 and E7 that inactivate the tumor suppressors P53 and Rb [56]. These mechanisms result in genomic instability, cell cycle dysregulation, and the reprogramming of cellular pathways that support stemness and immune evasion in HNSCC [57].

Significant differences in CSC marker expression are found between HPV-positive (HPV+) and HPV-negative (HPV−) HNSCC, suggesting distinct cancer cell subpopulations and therapeutic vulnerabilities. Studies have demonstrated that CSC markers such as CD44 and ALDH1 are differentially expressed based on HPV status. For example, CD44 expression is typically lower in HPV + cases compared to HPV− cases, correlating with worse outcomes [58,59]. Additionally, HPV-negative oropharyngeal cancer (OPC) CSCs show significantly higher pluripotency-associated genes such as POU5F1, SOX2, NANOG [13,60], Kruppler Like Factor 4 (KLF4) [61], and BMI1 [21,62] compared to HPV-positive tumors, indicating a more robust stem-like phenotype. Understanding these differences may guide the development of HPV+/specific targeted therapies.

5. Epigenetic modifiers and their role in HNSCC stem cells

Epigenetic modifiers are enzymes that regulate gene expression. They alter the chromatin structure and DNA accessibility, thereby modulating the transcriptional landscape across large genomic sections without changing the DNA sequence. Regulators that orchestrate both gene silencing and activation programs involved in maintenance, plasticity, and treatment resistance of CSCs in HNSCC are discussed here.

5.1. DNA methyltransferases (DNMTs)

DNA Methyltransferases (DNMTs) are enzymes that add methyl groups to DNA, leading to chromatin condensation and the silencing of differentiation-associated genes (Fig. 5A) [63]. In HNSCC, DNMT1 hypermethylates the promoters of p21 and p27, thereby promoting silencing of differentiation-related genes, contributing to CSC maintenance and poor prognosis [64].

Fig. 5. Epigenetic regulation of Cancer Stem Cells (CSCs) in HNSCC.

An illustration of chromatin modifiers in HNSCC CSCs. DNA Methyltransferases (DNMT1) induce chromatin condensation and silence differentiation genes. Histone Deacetylases (HDACs) (HDAC1, HDAC2) remove acetyl groups, causing further chromatin condensation and tumor suppressor gene repression. Histone Methyltransferases (HMTs) (EZH2) add methyl groups (H3K27me3) to histones, leading to gene silencing. Polycomb Group Proteins (BMI1) stabilize H3K27me3, reinforcing chromatin repression and promoting CSC self-renewal and tumor aggressiveness. On the other hand, Histone Acetyltransferases (HATs) (P300/CBP) add acetyl groups to histones, promoting transcription of stemness genes. These modifiers collectively drive tumor progression and resistance to therapies in HNSCC.

5.2. Histone Deacetylases (HDACs)

Histone Deacetylases (HDACs) such as HDAC1 and HDAC2, remove acetyl groups from histones, resulting in tighter chromatin and repression of gene expression (Fig. 5B). In CSCs, these enzymes repress the tumor suppressor genes P53 and RB through histone deacetylation, thereby enhancing CSC survival and contributing to therapy resistance [65].

5.3. Histone acetylation

Histone Acetylation on the other hand, promotes an open chromatin structure, facilitating the transcription of genes associated with stemness and plasticity [66] (Fig. 5C). Histone acetyltransferases (HATs) like P300/CBP catalyze this process and have been shown to enhance the expression of pluripotency factors [48,67].

5.4. Histone methyltransferases (HMTs)

Histone methyltransferases (HMTs) are enzymes that add methyl groups to specific histone residues, which can either activate or repress transcription depending on the context [68]. For example, the HMT EZH2, a key component of the Polycomb Repressive Complex 2 (PRC2), represses tumor suppressor genes by trimethylation of histone H3 at lysine 27 (H3K27me3), leading to chromatin compaction and reinforcement of CSC properties [69] (Fig. 5D).

5.5. Chromatin readers

In addition to writers and erasers, some chromatin readers are also involved in CSC regulation. They recognize specific histone modifications and help recruit or stabilize repressive complexes. One of the most extensively studied readers in HNSCC is BMI1, a core component of Polycomb Repressive Complex 1 (PRC1). BMI1 binds to H3K27me3 marks and further promotes chromatin condensation via histone H2A ubiquitination. Its expression correlates with CSC markers, treatment resistance, and poor clinical outcomes. Additionally, the activating methyltransferase SMYD3 is overexpressed in HNSCC, where it enhances stemness and EMT-related gene expression [70,71].

Together, epigenetic modifiers and readers coordinate a finely tuned regulatory system that dynamically represses differentiation pathways and activates stemness-associated gene programs, thereby contributing to CSCs’ plasticity, self-renewal, and therapeutic resistance.

6. HNSCC cancer stem cell metabolic flexibility

CSCs can adjust their metabolism in response to environmental conditions, a trait known as metabolic plasticity. Since the discovery of the Warburg effect a century ago, research has focused on the metabolic alterations that drive tumor progression. The bulk of the tumor cells rely on glycolysis for energy, even in the presence of oxygen [72]. However, CSCs can switch between glycolysis (Fig. 6A) and oxidative phosphorylation (OXPHOS) in mitochondria to generate ATP more efficiently (Fig. 6B). This adaptability allows CSCs to survive in different tumor regions, including areas with limited nutrient availability [73].

Fig. 6. Metabolic plasticity of Cancer Stem Cells (CSCs) in HNSCC.

This diagram illustrates the ability of CSCs to switch between A. glycolysis and B. oxidative phosphorylation (OXPHOS). C. Under low-oxygen or nutrient-limited conditions, HIF-1α activates PDK1, inhibiting the conversion of pyruvate into acetyl-CoA, thus promoting glycolysis over OXPHOS. D. When ATP levels further drop, AMPK activates mitochondrial biogenesis and enhances OXPHOS to support energy production. E. CSCs can also utilize glutamine through GLS1 for energy production via the TCA cycle. This metabolic flexibility contributes to CSC survival, resistance to therapy, and tumor progression.

One of the key regulators of this metabolic flexibility is Hypoxia-Inducible Factor 1-alpha (HIF-1α), a transcription factor commonly activated in low-oxygen (hypoxic) conditions [74]. HIF-1α promotes the expression of pyruvate dehydrogenase kinase 1 (PDK1), an enzyme that blocks the conversion of pyruvate (a glycolysis product) into acetyl-CoA (Fig. 6C), which is needed for OXPHOS [74]. Additionally, Peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) enhances mitochondrial function and oxidative metabolism, enabling resistance to oxidative stress and oxygen and nutrient fluctuations in the tumor microenvironment, and increases CSCs’ survival after chemotherapy or radiation exposure [75].

Conversely, AMP-activated protein kinase (AMPK) functions as an energy sensor that also helps CSCs survive under low-energy conditions [76] (Fig. 6D). When ATP (the cell’s main energy molecule) levels drop, AMPK activates mitochondrial biogenesis (the creation of new mitochondria) and promotes OXPHOS, allowing CSCs to produce energy when glucose is scarce [76,77]. In addition to glucose metabolism, CSCs can also use the amino acid glutamine as an alternative energy source. This is controlled by glutaminase 1 (GLS1) [47], which converts glutamine into glutamate that feeds into the tricarboxylic acid (TCA) cycle to sustain energy (Fig. 6D). Thus, CSCs can switch between glycolysis, OXPHOS, and other fuel sources, making them challenging to eliminate. Targeting these metabolic pathways may improve outcomes in HNSCC.

7. Resistance mechanisms of CSCs in HNSCC

Although CSCs are rare in tumors (Fig. 7A), following chemoradiation, the bulk of differentiated tumor cells die while CSCs persist (Fig. 7B). This survival advantage leads to further enrichment of CSCs with each treatment cycle. Their daughter cells retain these mechanisms, leading to enrichment of resistant clones (Fig. 7C).

Fig. 7. CSC-Associated Resistance Mechanisms in HNSCC.

A. CSCs contribute to treatment failure by resisting chemoradiation through multiple mechanisms. B. Tumor cells are eliminated, but CSCs persist. C. Surviving CSCs give rise to daughter cells that retain resistance, driving tumor regrowth. Key resistance mechanisms: D. Overexpression of ATP-binding cassette (ABC) transporters that pump out agents and reduce drug accumulation. E. Efficient DNA damage repair by activating both homologous recombination (HR) and non-homologous end joining (NHEJ) pathways. F. Quiescence, slow-cycling state, rendering them less susceptible to treatments targeting rapidly proliferating cells. G. Metabolic Flexibility: CSCs shift between glycolysis and oxidative phosphorylation to survive nutrient stress and maintain redox balance. These mechanisms enable clonal selection under therapeutic pressure and tumor relapse.

7.1. The increase of ATP-binding Glycoprotein cassette (ABC) transporters

A key mechanism by which CSCs resist chemotherapy is by efflux of cisplatin and 5-FU via ATP-binding cassette transporter (ABCG2), reducing intracellular drug accumulation and effectiveness [78] (Fig. 7D).

7.2. DNA damage repair (DDR) and CSC quiescence in response to treatment

Radiotherapy and chemotherapeutic agents used to treat HNSCC, such as cisplatin, carboplatin, paclitaxel, and 5-FU, induce DNA damage to eliminate rapidly dividing cells. But CSCs exhibit a robust DDR that facilitates efficient resolution of DNA lesions. Upregulated DDR proteins and enhanced homologous recombination (HR) and non-homologous end joining NHEJ contribute to resistance against double-strand breaks [48], allowing CSCs to evade apoptosis and DNA damage and persist post-treatment [79] (Fig. 7E). For example, POU5F1 activates PSMC3IP and RAD54L [80], while TRIP13 promotes error-prone NHEJ and DNA-PKcs, aiding resistance to double-strand breaks [81]. Furthermore, CSCs often replicate slowly and adopt a quiescent state (Fig. 7F), allowing them to evade treatments targeting replicating cells [13].

7.3. Metabolic reprogramming

As detailed in Paragraph 6, CSCs reprogram their metabolism through transcriptional and enzymatic regulators, which allows shifts in energy substrate utilization and resistance to metabolic stress (Fig. 7G). This metabolic flexibility also enables CSCs to resist oxidative damage caused by radiation or chemotherapy. Therefore, combinatorial therapeutic approaches aim to target multiple pathways simultaneously to enhance treatment outcomes.

8. Regulation of CSC gene expression by non-coding RNAs

Non-coding RNAs do not encode proteins but instead regulate mRNA stability and translation through multiple mechanisms. Specific ncRNAs could serve as biomarkers for CSCs, and targeting them may disrupt CSC functions, improving cancer detection and monitoring.

8.1. MicroRNAs

Post-Transcriptional Regulators of CSC Function and Therapeutic Targets: MicroRNAs (miRNAs) are 21–25 nucleotides in length and regulate genes by targeting mRNA for degradation or causing translational repression. They also bind to Toll-Like Receptors (TLRs), triggering pathways that promote inflammation, tumor growth, and metastasis. Dysregulated miRNAs can significantly impact CSC traits such as asymmetric division, tumorigenicity, and drug resistance. One of these miRNAs is miR-21, which promotes chemoresistance and tumor progression by downregulating the tumor suppressor PDCD4, upregulating the oncogenes BCL-2, MYC [82] and Inhibitor of Apoptosis Proteins (IAP), protecting CSCs from apoptosis. It also boosts the expression of stem cell genes like NANOG and STAT-3 signaling, which are important for CSC survival [83]. miR-34a acts as a tumor suppressor by targeting molecules like Flotillin-2 (FLOT-2) [84], Amphiregulin (AREG) [85], and MET [86], which regulate metastasis, proliferation, and cell migration in HNSCC. Upregulation of miR-302a/miR-302 b suppresses AOF1, AOF2 (histone demethylases), and DNMT1, promoting CSC self-renewal, clonal formation, and cisplatin resistance [15]. Conversely, dysregulation of miR-205 downregulates COMMD1, activating a feedback loop that increases inflammatory and stemness traits in HNSCC [87]. These miRNAs could be targeted for the treatment of HNSCC.

8.2. Long non-coding RNAs (lncRNAs): epigenetic modulators of CSC maintenance and tumor progression

LncRNAs are typically RNA molecules longer than 200 nucleotides [88] that regulate gene expression by modulating chromatin structure, affecting transcription, and interacting with various proteins and other RNAs to control numerous cellular processes. One notable example is Hox transcript antisense intergenic RNA (HOTAIR), which physically interacts with EZH2, a core component of the PRC2 complex, that methylates H3K27 to H3K27me3 and silences differentiation genes [89]. Another lncRNA upregulated in HNSCC is Metastasis-Associated Lung Adenocarcinoma Transcript 1 (MALAT1) [90]. MALAT1 promotes CSC phenotype, tumor growth, metastasis, and EMT [91]. Similarly, the upregulation of Linc-ROR is associated with CSC and tumor progression by regulating Forkhead box protein M1 (FOXM1) and LIM Domain Only 4 (LMO4). This activation activates the AKT/PI3K signaling pathway, increases cellular proliferation and invasion [92]. Lastly, the oncogenic lncRNA Taurine Upregulated Gene 1 (TUG1) is implicated in tumor progression and CSC maintenance in HNSCC. TUG1 promotes cisplatin resistance and enhances stemness properties by interacting with EZH2, resulting in repression of tumor suppressor genes involved in cell cycle control and apoptosis [93]. Taken together, lncRNAs contribute to CSC phenotype and therapy resistance by several physical interactions and epigenetic mechanisms.

8.3. Circular RNAs (circRNAs): miRNA sponges driving CSC traits and tumor progression

CircRNAs are another class of non-coding RNAs with a covalently closed-loop structure, which confers high stability and resistance to exonuclease degradation [94]. Studies suggest that they function as miRNA sponges, thus reducing their inhibitory effects and indirectly regulating target gene expression [95]. In HNSCC, multiple circRNAs are important to the CSC phenotype. circSHKBP1 increases tumor growth by sequestering miR-766–5p, thereby increasing the expression of BMI1 [96]. Additionally, circFAT1 was shown to enhance CSC traits by activating STAT3. Other circRNAs, such as circ_0000045 [97], circPVT1 [98], and circ-CCND1 [99], regulate CSC traits and contribute to tumor progression. Targeting circRNAs may eliminate CSCs and enhance the effectiveness of treatments in HNSCC.

9. Immune evasion mechanisms of HNSCC CSCs

CSCs create an immune-excluded microenvironment around them, both directly and indirectly (Fig. 8A), also known as “immune deserts” [100]. They secrete cytokines such as TGF-β and IL-10, which recruit myeloid-derived suppressor cells (MDSCs) and promote regulatory T-regulatory (Treg) cell activation (Fig. 8B).

Fig. 8. CSCs affect immune responses and create an immune-excluded niche.

A. CSC-rich tumor areas often become immune-excluded: B. The innate immune response: They release cytokines (such as IL-10 and TGF-β), which reduce the effectiveness of natural killer (NK) cells and increase the number of myeloid-derived suppressor cells (MDSCs). They also activate regulatory T cells (Tregs), which further weaken the innate immune response. C. Adaptive immune response: CSCs express low levels of MHC-I, impairing antigen presentation to CD8⁺ T cells, while upregulating immune checkpoint ligands such as PD-L1; this deactivates CD8⁺ cytotoxic T cells, leading to T-cell exhaustion. Common to both responses is Treg activation, which further suppresses CD8⁺ T cells and blocks CD4⁺ T-cell activation through CTLA-4 binding to CD80/CD86 on antigen-presenting cells.

9.1. The innate immune response

Innate immune-modulatory signals from CSCs suppress cytotoxic T-cell (CD8⁺) recruitment and dampen CD4⁺T-cell and Natural killer (NK) cell responses. They also suppress Myeloid cells, a broad group of innate immune cells derived from bone marrow myeloid progenitors. This group includes monocytes (which give rise to macrophages and dendritic cells), granulocytes (neutrophils, eosinophils, and basophils), mast cells, and myeloid-derived suppressor cells (MDSCs). This is another important feedback loop since MDSCs further suppress both cytotoxic T and NK cell responses, promote regulatory T-cell expansion, and sustain an immunosuppressive cytokine that protects CSCs from immune clearance and supports their maintenance (Fig. 8B). At the same time, CSCs downregulate major histocompatibility complex (MHC) class I molecules, reducing their recognition by cytotoxic CD8⁺ T cells [101]. Together, these mechanisms impair immune surveillance, allowing CSCs to evade destruction.

9.2. The adaptive immune response modulation

CSCs also modulate the adaptive immune response by upregulating checkpoint molecules such as programmed death-ligand 1 (PD-L1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) [102] (Fig. 8C). PD-L1 on CSCs binds PD-1 on T cells, inducing T-cell anergy and apoptosis. In HNSCC, these immune evasion mechanisms have been confirmed in patient samples. A cohort of 21 primary human HNSCC tumors and patient-derived xenografts, CD44-positive CSCs showed higher PD-L1 expression by both flow cytometry and qRT-PCR (Combined Positive Score≥ 1 or ≥ 5) and reduced immunogenicity [103].

CSCs also shape an immunosuppressive niche by recruiting Treg cells that express CTLA-4, which binds to CD80/CD86 on antigen-presenting cells (Fig. B, C), inhibiting CD4⁺ T cell activation and further dampening immune responses [104]. These checkpoint defenses, combined with CSC-driven cytokine signaling, enable immune escape, promote therapeutic resistance, and drive HNSCC progression and recurrence.

10. Evasion of disulfidptosis: a new mechanism and a potential target for drug design

One of the defining traits of CSCs is their enhanced ability to survive under therapeutic and Metabolic stress. While cell stress commonly triggers cell death through apoptosis [105], necroptosis [106], or ferroptosis [107], these mechanisms do not fully explain the cytoskeletal collapse observed under glucose starvation. A newly characterized form of cell death, called disulfidptosis, is triggered when excessive cystine uptake leads to disulfide-bond accumulation in actin filaments, causing cytoskeletal failure. Unlike other forms of cell death, it is driven by cystine-induced disulfide stress, and not by ROS or lipid damage [108].

HNSCC CSCs often express high levels of SLC7A11 and its chaperone CD98hc (SLC3A2), which localize it to the plasma membrane [109]. This cystine transporter facilitates extracellular cystine uptake to support glutathione (GSH) synthesis and maintain redox balance, especially under stress conditions like glucose deprivation [110]. This enables efficient cystine import and supports antioxidant defenses. CSCs also upregulate detox enzymes such as GPX4 and ALDH1A1, and maintain high GSH levels, creating a redox-protected niche that helps them survive where other tumor cells may die. Ironically, this reliance on SLC7A11 makes CSCs vulnerable to disulfidptosis, especially under glucose-deprived conditions. Thus, targeting the SLC7A11-CD98hc axis may provide a novel strategy to eliminate CSCs in HNSCC through disulfide-driven cell death.

11. Emerging precision therapies targeting CSCs in HNSCC

11.1. CSC targeting drugs

Although many molecular pathways protect CSCs from conventional therapies, they can also serve as potential drug targets. Several small-molecule inhibitors have been identified and developed, with many candidates entering clinical trials. Inhibition of CSC pathways may deplete their self-renewal and metastatic capabilities. Here are some examples of small-molecule inhibitors designed to target CSCs.

11.1.1. JAK/STAT inhibitors

Clinical and preclinical evidence support this pathway as a therapeutic target in HNSCC. The JAK1/2 inhibitor Ruxolitinib reduces STAT3 phosphorylation and NANOG expression in patient-derived xenograft models [111], and is currently under investigation in phase I clinical trials [112].

11.1.2. Wnt/β-catenin-specific inhibitors

While none are currently FDA-approved for cancer, several agents, including repurposed drugs, show promise. For example, PORCN inhibitors block the secretion of Wnt ligands by inhibiting their palmitoylation, thereby preventing activation of the Wnt/β-catenin pathway and disrupting CSC signaling [113]. The PORCN inhibitor WNT974 (LGK974) reduces tumor growth in HNSCC models and is in a Phase II trial that includes HNSCC patients [113].

11.1.3. Inhibitors targeting ALDH1A1

Disulfiram, an ALDH1A1 inhibitor, sensitizes CSCs to cisplatin and radiation in a preclinical xenograft model of HNSCC [114]. Disulfiram combined with copper induces ROS and disrupts pro-survival pathways, leading to CSC death and reduced tumor growth [115]. Suppressing CSC-like self-renewal and cisplatin resistance supports ALDH1A1 as a functional target in HNSCC CSC populations.

11.1.4. Inhibition of Epigenetic Enzymes

Inhibition of Epigenetic Enzymes like HDACs and EZH2 induces p21 and p27 expression, which mediate G0 phase cell cycle arrest and restore differentiation programs silenced in CSCs [116].

11.2. Emerging Precision Immunotherapies Targeting Stemness in HNSCC in preclinical and clinical trials

Emerging Precision Immunotherapies Targeting Stemness in HNSCC in Preclinical and Clinical Trials Precision medicine tailored to specific characteristics of patient tumors is heavily studied, with emerging treatments beginning to enter clinical use. As research continues to identify more pathways involved in HNSCC, new targeted therapies are expected to become available. Here are some effective immunotherapies being explored and used for HNSCC treatment.

11.2.1. Immune checkpoint inhibitors targeting cancer stemness

Immune checkpoint inhibitors show promise as a therapeutic strategy to improve treatment outcomes in advanced HNSCC. Small molecules and peptides that block PD-1, PD-L1, or CTLA-4 are in preclinical-stage research. Furthermore, combining PD-1 and CTLA-4 blockers may be a better therapeutic strategy for CSC-driven immune suppression.

11.2.2. Therapeutic monoclonal antibodies

Cetuximab is an FDA-approved anti-EGFR monoclonal antibody used in HNSCC [117]. It works by blocking EGFR signaling, which is often overexpressed in HNSCC tumors. Cetuximab inhibits PI3K/AKT and MAPK pathways and indirectly reduces CSC populations [118].

Other antibodies block PD-1 to enhance CD8⁺ T-cell cytotoxicity against CSCs [119]. By blocking the PD-1/PD-L1 axis, they strengthen the body’s immune response, improving treatment efficacy and potentially reducing recurrence by targeting the root of tumorigenesis. Nivolumab is one of the PD-1 inhibitors approved by the FDA for the treatment of recurrent or metastatic HNSCC following the CheckMate-141 clinical trial [120]. Another FDA-approved anti-PD-1 antibody is Pembrolizumab, for non-resectable or metastatic HNSCC. In the KEYNOTE-048 trial, Pembrolizumab has improved overall survival compared to cetuximab when used alone or in combination with chemotherapy, in patients with high PD-L1 expression [121].

More recently, the FDA approved a treatment protocol for resectable locally advanced HNSCC, based on the results of the KEYNOTE-689 clinical trial. In this protocol, Pembrolizumab is used before surgery to shrink the tumor, activate immune responses, and reduce micrometastasis. It is then continued after surgery to eliminate residual cancer cells and prevent recurrence [122].

11.2.3. Antibody-drug conjugate (ADC)

CD44v6 (CD44 variant isoform 6) is another example of a membrane marker enriched in a particularly aggressive subpopulation of CSCs [123]. Studies have shown that Bivatuzumab mertansine, an antibody-drug conjugate (ADC) targeting CD44v6, not only reduces the CSC population but also disrupts the tumor microenvironment [124]. However, clinical development of bivatuzumab mertansine was halted due to dose-limiting toxicity, including a fatal case of toxic epidermal necrolysis, despite signs of modest but promising clinical activity in early trials [125].

Trastuzumab deruxtecan (T-DXd) is another FDA-approved antibody-drug conjugate targeting HER2 in solid tumors, including HNSCC tumors with high HER2 expression [126]. Ongoing clinical trials support its use as a promising therapeutic strategy in HER2-positive or HER2-expressing HNSCC tumors [127].

11.3. CAR-T cell therapy targeting CSC-associated antigens in HNSCC

Checkpoint inhibitors such as PD-1/PD-L1 and CTLA-4 blockade have shown promise in subsets of HNSCC patients, but many tumors remain resistant, in part due to the immunosuppressive niche created by CSCs. As CSCs actively remodel the tumor microenvironment to escape immune recognition, novel immunotherapeutic strategies are being studied. For example, Chimeric antigen receptor (CAR) T cells are T lymphocytes engineered to express synthetic receptors that recognize and kill tumor cells based on surface antigens, independent of MHC presentation. CAR-T cells targeting the splice variant of CD44, CD44v6, demonstrated selective killing of tumor cells in vitro and durable tumor suppression in xenograft models, with minimal toxicity to normal tissues [128]. These findings suggest that CAR-T cells could eliminate therapy-resistant CSCs, offering a precision-based approach to prevent relapse. While clinical validation remains limited, ongoing trials and next-generation CAR aimed at enhancing tumor infiltration and persistence may expand the therapeutic effect in HNSCC.

11.4. Combinatorial therapeutic strategies targeting CSCs

Recent advances in HNSCC aim to target CSCs and overcome immune evasion or therapy resistance. Combining chemoradiation or targeted agents with NK cell-based immunotherapy significantly enhances antitumor responses in CSC-rich tumors [129].

Similarly, targeting the DDR pathways sensitizes CSCs to chemoradiation and prevents repair-driven resistance [130]. Combining radiotherapy with PARP inhibitors or NHEJ inhibitors has shown potential to suppress CSCs in preclinical HNSCC models [79].

Near-infrared photoimmunotherapy combined with pembrolizumab achieved a 100 % overall response rate, with three remaining disease-free after one year [131].

Inhibition of Aurora kinase A enhances the efficacy of checkpoint inhibitors by reducing CSC-mediated immune suppression [132]. These combinatorial strategies validate the potential of disrupting CSC survival networks while simultaneously restoring immune recognition via MHC-I upregulation and PD-L1 blockade, reducing tumor recurrence [101].

12. Conclusions and prospects

CSCs are central to the development and therapeutic resistance of HNSCC tumors. These cells have unique properties, including quiescence, efficient DNA repair, and telomere maintenance, which allow them to survive common cancer treatments with unlimited proliferative potential. This review outlines how these diverse signaling pathways, metabolic adaptations, immune evasion strategies, and epigenetic reprogramming converge to preserve the CSC phenotype.

While integrating both mechanistic and emerging clinical evidence, long-term validation of CSC-directed therapies in HNSCC patients remains limited. Current studies rely on preclinical models, including in vitro assays and xenografts, which may not fully recapitulate the tumor complexity in patients. Moreover, heterogeneity in CSC markers, isolation techniques, and functional assays across studies presents challenges for clinical translation. To address this, we highlighted both supportive and conflicting clinical data throughout the review.

Future progress may require standardized definitions of CSC markers using methods such as single-cell multi-omics and functional lineage tracing to monitor their functional behavior over time. Robust randomized clinical trials are also needed to translate CSC-targeted strategies into durable therapeutic outcomes. These tools can improve target validation and accelerate the clinical translation of CSC-directed therapies in HNSCC.

To bridge the gap between current and future treatments, it is essential to develop targeting approaches that address the intricacies of CSC mechanisms. Promising therapeutic approaches under development include inhibitors of stemness-associated pathways such as Wnt, Notch, STAT3, epigenetic modulators, and agents that reverse CSC quiescence or metabolic reprogramming. Novel immunotherapies, including CSC-directed CAR-T-cells and immune checkpoint inhibitors, aim to restore anti-tumor immunity and eliminate immune-privileged CSCs. In addition, non-coding RNAs (ncRNAs) have emerged as regulators of CSC maintenance and therapy resistance, offering opportunities for RNA-based interventions and biomarker development. A deeper understanding of how these axes intersect could enable the development of precision-based therapies capable of eliminating CSCs and improving durable responses in HNSCC.

Current standard-of-care treatments following surgery, such as radiation or chemotherapy, often fail to eradicate this subpopulation in advanced HNSCCs. Moving forward, a major research goal is to develop integrated treatments that combine CSC-targeted therapies to prevent cancer relapses and metastasis. New therapies may avoid off-target toxicity in normal somatic stem cells by targeting CSC-specific markers [133]. Future studies should prioritize identifying CSC vulnerabilities, understanding tumor heterogeneity, and designing delivery systems that effectively target CSCs within the tumor. Ultimately, eliminating self-renewal, telomere maintenance, and immune evasion could render tumors incapable of indefinite growth, transforming them into benign, non-invasive-like lesions. Therefore, targeting CSCs may offer a path not only to remission but to durable cures in HNSCC.

Abbreviation

ABC

ATP-Binding Cassette

ADC

Antibody-Drug Conjugate

ALDH1A1

Aldehyde Dehydrogenase 1 Family Member A1

AMPK

AMP-Activated Protein Kinase

ARE

Amphiregulin

BMI1

B lymphoma Mo-MLV Insertion Region 1

CAR-T

Chimeric Antigen Receptor T cell

CD44v6

CD44 Variant Isoform 6

circRNA

Circular RNA

CSC

Cancer Stem Cell

CTGF

Connective Tissue Growth Factor

CTLA-4

Cytotoxic T-Lymphocyte Associated Protein 4

DDR

DNA Damage Repair

DNMT

DNA Methyltransferase

EMT

Epithelial-to-Mesenchymal Transition

EpCAM

Epithelial Cell Adhesion Molecule

EZH2

Enhancer of Zeste Homolog 2

FLOT-2

Flotillin-2

GLS1

Glutaminase 1

GSH

Glutathione

HAT

Histone Acetyltransferase

HDAC

Histone Deacetylase

HER2

Human Epidermal Growth Factor Receptor 2

HIF-1α

Hypoxia-Inducible Factor 1 Alpha

HMT

Histone Methyltransferase

HPV

Human Papillomavirus

HNSCC

Head and Neck Squamous Cell Carcinoma

IL-10

Interleukin 10

JAK

Janus Kinase

KLF4

Kruppel-Like Factor 4

LGR5

Leucine-rich Repeat-containing G-Protein Coupled Receptor 5

lncRNA

Long Non-Coding RNA

MALAT1

Metastasis-Associated Lung Adenocarcinoma Transcript 1

MAPK

Mitogen-Activated Protein Kinase

MDSC

Myeloid-Derived Suppressor Cell

miRNA

MicroRNA

MMP

Matrix Metalloproteinase

MHC-I

Major Histocompatibility Complex Class I

MYC

Myelocytomatosis Oncogene

NADPH

Nicotinamide Adenine Dinucleotide Phosphate (Reduced)

NANOG

Homeobox Protein NANOG

NK

Natural Killer (Cell)

OCT3/4

Octamer-Binding Transcription Factor 3/4 (POU5F1)

OXPHOS

Oxidative Phosphorylation

PARP

Poly (ADP-ribose) Polymerase

PD-1

Programmed Cell Death Protein 1

PD-L1

Programmed Death-Ligand 1

PDK1

Pyruvate Dehydrogenase Kinase 1

PGC-1α

Peroxisome Proliferator-Activated Receptor Gamma Coactivator 1-alpha

PI3K

Phosphoinositide 3-Kinase

PRC1/2

Polycomb Repressive Complex ½

PSMC3IP

Proteasome 26S Subunit, ATPase 3 Interacting Protein

PTCH

Patched (Hedgehog receptor

RB

Retinoblastoma Protein

ROS

Reactive Oxygen Species

SHH

Sonic Hedgehog

SMO

Smoothened (receptor)

SOX2

SRY-box Transcription Factor 2

STAT3

Signal Transducer and Activator of Transcription 3

SUFU

Suppressor of Fused

TCA

Tricarboxylic Acid (Cycle

TERT

Telomerase Reverse Transcriptase

TGF-β

Transforming Growth Factor Beta

TNM

Tumor Node Metastasis

TUG1

Taurine Upregulated Gene 1

TWIST

Twist Family BHLH Transcription Factor

ZEB1/ZEB2

Zinc Finger E-Box Binding Homeobox 1/2

ZSCAN4

Zinc Finger and SCAN Domain Containing 4