Targeting cancer stem cell plasticity and tumor microenvironment crosstalk: a comprehensive review

Abstract

Cancer remains a major global health challenge, marked by profound heterogeneity and adaptability that frequently limit the success of conventional and targeted therapies. Among the diverse cellular populations within tumors, cancer stem cells (CSCs) have emerged as key drivers of initiation, progression, metastasis, and therapeutic resistance. Sharing core features with normal stem cells, including self-renewal and multipotency, CSCs sustain tumor growth and contribute to intratumoral heterogeneity. Importantly, the traditional hierarchical model has been revised with the recognition of CSC plasticity—the dynamic ability of cells to transition between CSC and non-CSC states or among distinct CSC subsets. This phenotypic flexibility, governed by intrinsic genetic and epigenetic mechanisms and strongly influenced by extrinsic cues from the tumor microenvironment (TME), underpins therapeutic resistance, immune evasion, and disease recurrence. The TME represents a complex, dynamic ecosystem composed of stromal cells, immune populations, extracellular matrix components, and soluble factors. Far from a passive scaffold, the TME actively engages in reciprocal crosstalk with CSCs, sustaining stemness, modulating plasticity, and shaping tumor progression. This bidirectional interplay reinforces CSC survival, fuels metastatic potential, and contributes to therapeutic failure. A comprehensive understanding of CSC plasticity and CSC–TME interactions is therefore essential for the design of durable therapeutic strategies. This review synthesizes current knowledge of the molecular and cellular mechanisms driving CSC plasticity and their microenvironmental regulation. It further evaluates emerging approaches—including small molecules, epigenetic modulators, and TME-normalizing therapies—that target these processes, and discusses challenges and future directions for translating these strategies into precision oncology.

Introduction

Cancer remains a formidable global health challenge, characterized by its remarkable heterogeneity and intrinsic adaptability, which frequently undermine the efficacy of conventional and targeted therapies alike [1–5]. This inherent complexity stems from diverse cellular populations within tumors, including a critical subset known as cancer stem cells (CSCs). CSCs, a small subpopulation of cells within tumors, are increasingly recognized as the principal drivers of tumor initiation, progression, metastasis, and therapeutic resistance, exhibiting properties akin to normal stem cells such as self-renewal and multipotency [6, 7]. Their ability to self-renew and differentiate into various cancer cell phenotypes contributes significantly to intratumoral heterogeneity and the maintenance of tumor populations [6, 8]. However, the traditional hierarchical CSC model has evolved with the recognition of profound CSC plasticity—the dynamic ability of cancer cells to interconvert between CSC and non-CSC states, or among different CSC subsets [9]. This phenotypic flexibility, driven by both intrinsic genetic and epigenetic factors, and profoundly shaped by non-cell-autonomous effects from the tumor microenvironment (TME), enables rapid adaptation to therapeutic pressures and facilitates immune evasion, thus fueling disease recurrence and progression [9, 10].

The tumor microenvironment, a complex and dynamic ecosystem surrounding cancer cells, comprises a heterogeneous array of cellular components, including stromal cells (e.g., cancer-associated fibroblasts, endothelial cells, mesenchymal stem cells), diverse immune cells (e.g., tumor-associated macrophages, T lymphocytes, natural killer cells, myeloid-derived suppressor cells), and non-cellular elements such as the extracellular matrix and a rich milieu of soluble factors (cytokines, chemokines, growth factors, metabolites) [11, 12]. This intricate TME is not merely a passive bystander but an active participant, engaging in critical, bidirectional crosstalk with CSCs. This symbiotic relationship profoundly influences CSC fate, sustaining their stemness, promoting their survival, dictating their phenotypic plasticity, and ultimately dictating tumor growth, metastasis, and resistance to therapy [13, 14].

While numerous reviews have addressed either CSC plasticity or the TME’s role in cancer, the novelty of this review lies in its comprehensive synthesis of these two dynamic systems as a single, co-dependent, and therapeutically-targetable axis. We move beyond a parallel description to build a conceptual framework where intrinsic cellular plasticity (e.g., EMT, dedifferentiation, metabolic reprogramming) is inextricably linked to its extrinsic microenvironmental regulation (e.g., signals from CAFs, TAMs, and the vascular niche). A unique focus of this work is the critical evaluation of emerging therapeutic strategies that specifically target this interface, including a dedicated analysis of novel epigenetic modulators and the burgeoning field of m6A RNA modification inhibitors. By connecting the molecular mechanisms of plasticity directly to their TME drivers and, in turn, to the next generation of translational strategies, this review provides a distinct, clinically-focused roadmap for precision oncology.

Cancer stem cell plasticity: mechanisms and functional consequences

Defining the dynamic nature of cancer stem cells

Cancer stem cells (CSCs) are a distinct subpopulation within tumors endowed with self-renewal and multipotency, which are foundational to tumorigenesis, metastasis, and therapeutic resistance [6, 7, 15]. First identified in leukemia (CD34+/CD38−) and now isolated from solid tumors like breast cancer (CD44 + CD24-/low), their ability to drive tumor initiation and heterogeneity makes them a critical therapeutic target [6, 16].

However, the traditional hierarchical model has been replaced by the concept of “CSC plasticity”—the dynamic ability of cells to interconvert between CSC and non-CSC states or switch among different CSC subsets [9, 17]. This phenotypic flexibility, crucial for adapting to therapeutic pressures, is a primary driver of resistance [18, 19]. Compounding this, many CSCs adopt a quiescent (G0) state, lacking active replication and metabolic activity, which shields them from conventional treatments [15]. Eliminating proliferative cells can inadvertently stimulate these dormant CSCs to reactivate, creating a critical therapeutic dilemma that necessitates targeting both plasticity and quiescence [15, 20].

Cellular mechanisms of phenotypic plasticity: epithelial-mesenchymal transition (EMT), mesenchymal-epithelial transition (MET), and dedifferentiation/reprogramming

The dynamic nature of cancer cells, particularly their ability to switch phenotypes, is fundamental to cancer initiation, progression, and therapeutic evasion [21]. This adaptability is largely mediated by cellular processes such as Epithelial-Mesenchymal Transition (EMT), its reverse, Mesenchymal-Epithelial Transition (MET), and cellular dedifferentiation or reprogramming [22].

Epithelial-mesenchymal transition (EMT)

EMT is a highly dynamic and reversible biological process fundamental to embryonic development and wound healing, which is aberrantly co-opted in cancer progression [23]. During EMT, epithelial cells shed their apical-basal polarity and cell-cell junctions, acquiring a mesenchymal phenotype characterized by increased motility, invasiveness, and resistance to programmed cell death [23, 24]. This transformation is predominantly driven by epigenetic modifications rather than stable genetic alterations. Key regulators include EMT-inducing transcription factors (EMT-TFs) like Snail, Slug, Zeb1/Zeb2, and Twist, as well as specific microRNAs (miRs). Signaling pathways such as transforming growth factor (TGF)-β, Wingless/Integrated (WNT), Notch, and Hippo also contribute significantly [17, 25, 26]. Critically, EMT is closely associated with the acquisition of both invasive and stem-cell properties in cancer cells, and its initiation can generate a CSC-like population, thereby promoting tumor recurrence and metastatic dissemination [23]. For instance, in breast cancer, overexpression of Twist, Snail, or FOXC2 not only imparts mesenchymal properties but also increases the expression of CD44+/CD24-/low breast CSC markers and mammosphere-forming efficiency [27]. Similarly, in prostate cancer, cells with an EMT phenotype show increased expression of Sox2, Nanog, Pou5F1, lin28B, and/or Notch1, along with enhanced sphere-forming ability [27] (Fig. 1). This connection between EMT and therapeutic resistance is a critical clinical target. In nasopharyngeal carcinoma, for example, radioresistance is strongly associated with the EMT program. A recent study by Huang et al. demonstrated that the tyrosine kinase inhibitor (TKI) afatinib could promote radiosensitivity specifically by reversing EMT. Mechanistically, this was achieved by inhibiting the CD44-Stat3 signaling axis, providing a direct link between a stemness marker (CD44), a core signaling pathway (STAT3), and the EMT-driven resistance phenotype [28]. Clinically, this direct link between EMT and stemness suggests that therapeutic strategies aimed at inhibiting key EMT-TFs or their upstream signaling pathways (e.g., TGF-β, WNT) could represent a powerful approach to block metastatic dissemination and help overcome drug resistance.

Fig. 1.

Diagrammatic representation of epithelial-to-mesenchymal transition (EMT). Epithelial cells with apical-basal polarity and adherens junctions lose polarity and intercellular adhesion under the influence of EMT-inducing transcription factors (Snail, Slug, Zeb1/2, Twist), microRNAs, and key signaling pathways (TGF-β, WNT, Notch, Hippo). This transition drives the acquisition of a mesenchymal phenotype characterized by enhanced motility, invasiveness, and resistance to apoptosis, ultimately contributing to cancer progression, metastasis, and CSC-like properties [29]

Mesenchymal-epithelial transition (MET)

MET represents the reverse process of EMT, where mesenchymal cells revert to an epithelial phenotype [30]. This transition is crucial in embryogenesis and organogenesis, and its dysregulation in cancer can lead to a more mesenchymal phenotype that is either reversible or not [31]. While EMT facilitates invasion and dissemination, MET is often required for metastatic colonization at secondary sites, allowing disseminated mesenchymal cells to re-acquire epithelial characteristics necessary for sustained growth within a new tissue [32]. Targeting MET is being investigated as a therapeutic strategy to reverse the mesenchymal phenotype to an epithelial one, potentially re-sensitizing cells to therapy [33]. This provides a key clinical lever: pharmacologically inducing MET in disseminated, mesenchymal-like tumor cells could theoretically ‘trap’ them in a less invasive state at secondary sites, thereby inhibiting metastatic colonization and restoring sensitivity to epithelial-targeting therapies.

Dedifferentiation and reprogramming

Cancer cells exhibit a remarkable capacity for dedifferentiation, a process where a specialized cell type loses its lineage-specific gene expression and adopts a more primitive, stem-like signature [34]. This phenomenon bears striking resemblance to the generation of induced pluripotent stem cells (iPSCs), where differentiated somatic cells are reprogrammed back to a pluripotent state with unlimited proliferative properties and self-renewal capacity [34]. Oncogenic transformation often involves the de novo acquisition of developmental programs, similar to cellular reprogramming, resulting in cells with unlimited self-renewal potential—a hallmark shared with iPSCs [34]. This dedifferentiation can be driven by epigenetic disruption, even in the absence of genetic mutations, playing a significant role in cancer development and plasticity [35]. For example, a study using in vivo reprogrammable mice showed that incomplete reprogramming could lead to tumors resembling Wilms’ tumor, primarily through epigenetic disruption rather than genetic mutations [35]. Factors such as p53 inhibition, hTERT activation, and aberrant acquisition of NODAL, NOTCH, and WNT proteins facilitate this plasticity [36]. The ability of non-CSCs to dedifferentiate and acquire CSC-like properties under certain conditions challenges the strict unidirectional hierarchical CSC model, contributing to tumor heterogeneity [34]. From a clinical standpoint, this reprogramming capacity is a critical vulnerability; “differentiation therapies” that force CSCs to exit their stem-like state (e.g., using epigenetic modulators or retinoids) represent a tangible strategy to exhaust the self-renewing tumor population and prevent relapse.

Phenotype switching

This term encompasses the dynamic and often reversible changes in cellular states that profoundly influence drug receptivity and metastatic potential. Such switching can occur spontaneously during tumor development in a three-dimensional environment or be induced by therapeutic interventions [37, 38]. It is governed by a complex interplay of both cell-intrinsic factors (e.g., oncogenic mutations) and cell-extrinsic factors (e.g., microenvironmental cues). For instance, cancer cells can strategically shift between a drug-sensitive, proliferative state and a drug-tolerant, slow-cycling “persister” state, enabling them to endure lethal drug concentrations through transient, reversible changes [17, 39]. This adaptability is a critical attribute of cancer biology, allowing cells to rapidly alter their phenotypic and molecular profiles in response to environmental pressures like drug exposure.

Molecular drivers of CSC plasticity

The self-renewal, differentiation, proliferation, invasion, migration, and therapeutic resistance of CSCs are intricately regulated by a complex network of intracellular signaling pathways, epigenetic modifications, and non-coding RNAs, all of which contribute to their remarkable plasticity.

Key signaling pathways

Numerous signaling pathways are aberrantly activated or dysregulated in CSCs, driving their stemness and adaptive capabilities [6]. These pathways often exhibit significant crosstalk, contributing to the robustness of CSC phenotypes and their resistance to targeted therapies (Table 1).

Table 1.

Different signaling pathways, their related mechanisms, and role in CSC

Pathway	General role in CSC plasticity and tme crosstalk	Representative cancer types	Therapeutic examples (lead agents)	References
Wnt/β-Catenin	Promotes CSC proliferation, self-renewal, EMT, and immune evasion; forms feedback loops with TME factors.	Colorectal, Pancreatic, Ovarian, AML	Porcupine Inhibitors: LGK974β-catenin Inhibitors: PRI-724Fzd Antagonists: Ipafricept	[40–43]
Notch	Governs cell fate, stem cell maintenance, and tumor heterogeneity. Aberrant activation is common.	Breast (TNBC), HNSCC, T-cell ALL, Desmoid Tumors	γ-Secretase Inhibitors (GSIs): Nirogacestat, MK-0752DLL4 Antibodies: Demcizumab	[42, 44, 45]
NF-κB	Central to inflammation and immunity; constitutively activated in CSCs, promoting survival, anti-apoptosis, EMT, and chemoresistance.	Pancreatic, Breast, Multiple Myeloma	Inhibitors: Parthenolide, Triptolide(Also targeted by proteasome inhibitors)	[46–48]
JAK/STAT	Constitutively activated by TME factors (e.g., IL-6); promotes stemness, proliferation, and therapy resistance.	Colorectal, Pancreatic, HNSCC	STAT3 Inhibitors: Napabucasin (BBI608)JAK Inhibitors: Ruxolitinib	[45, 49]
TGF-β	Dual role; acts as an oncogene in advanced cancer, inducing EMT, immune suppression, and therapeutic resistance.	Pancreatic (PDAC), Glioblastoma (GBM), Hepatocellular (HCC)	TGF-βR1 Inhibitors: GalunisertibLigand Traps: Bintrafusp alfa	[41, 50–52]
PI3K/AKT/mTOR	Drives cell survival, growth, and proliferation; frequently activated (e.g., PIK3CA mutation, PTEN loss) and linked to resistance.	Glioblastoma (GBM), Breast (ER+), Prostate	PI3K Inhibitors: Alpelisib, BuparlisibBrain-penetrant: GDC-0084	[45, 53, 54]
MAPK	Regulates proliferation, survival, and differentiation. The p38 MAPK arm is specifically involved in CSC development, metastasis, and chemoresistance.	Melanoma (BRAF-mutant), NSCLC, Colorectal	p38 Inhibitors: (e.g., Ralimetinib)MEK Inhibitors: TrametinibBRAF Inhibitors: Vemurafenib	[41, 55, 56]
Hedgehog (Hh)	Aberrantly activated in malignancies; influences tumor formation, progression, drug resistance, and CSC maintenance.	Basal Cell Carcinoma (BCC), Medulloblastoma, AML	SMO Inhibitors: Vismodegib, Sonidegib, GlasdegibGLI Inhibitors: GANT-61	[42, 45, 57, 58]

Epigenetic regulation

Epigenetic modifications, defined as heritable changes in gene expression not involving DNA sequence alterations, are pivotal conductors of cellular plasticity. Dysregulation in epigenetic control, encompassing DNA methylation, histone modifications, and chromatin remodeling, can lead to stalled differentiation or aberrant cell reprogramming, contributing to malignant transformation [59, 60]. These changes often precede and outnumber genetic aberrations, highlighting their profound impact [61].

Aberrant DNA methylation patterns are central to epigenetic changes in cancer, capable of silencing tumor suppressor genes and activating oncogenes [61]. Sweeping hypomethylation across the genome is a hallmark of cancer, promoting instability, while targeted hypermethylation at CpG islands (CGIs) in gene promoters silences tumor suppressor genes like RB1, CDKN2, and MLH1 [61]. Studies have shown that lower levels of DNA methylation correlate with high global plasticity in glioblastoma, while IDH1 mutations promoting hypermethylation restrict plasticity [18, 62].

Histone modifications directly influence chromatin structure and gene expression. EMT-inducing transcription factors (EMT-TFs) recruit specific chromatin-modifying enzymes to effect widespread changes in gene expression during EMTs [63]. For example, the epigenetic silencing of E-cadherin, a hallmark of EMT, involves Polycomb group (PcG) proteins (PRC1 and PRC2 complexes) and histone deacetylases (HDACs) [63]. SNAIL, an EMT-TF, physically interacts with EZH2 and SUZ12 at the.

CDH1 promoter to catalyze H3K27 trimethylation, silencing CDH1 transcription [63]. The coexistence of both repressive (H3K27me3) and activating (H3K4me3) modifications on certain promoters, known as a ‘bivalent’ state, allows dynamic regulation of gene expression and contributes to the plasticity of mammary epithelial cells [63]. Chromatin remodeling complexes, particularly the SWI/SNF family, also play crucial roles in DNA damage response and are frequently mutated in cancer, impacting chromatin accessibility and gene expression [18, 19]. The functional dynamics of ATP-dependent chromatin remodelers are subject to modulation by both extracellular and intracellular cues [61]. These epigenetic changes are heavily influenced by environmental factors, such as aging, carcinogens, and the tumor microenvironment itself, acting as a bridge between the environment and the epigenome [61].

Non-coding RNAs (miRNAs, lncRNAs, exosomes)

Non-coding RNAs (ncRNAs), including microRNAs (miRNAs) and long non-coding RNAs (lncRNAs), and extracellular vesicles like exosomes, are crucial regulators of gene expression and play significant roles in mediating CSC plasticity and TME crosstalk.

MiRNAs are short noncoding RNA fragments that regulate gene expression post-transcriptionally by binding to target mRNAs, impeding translation or prompting mRNA destabilization. Dysregulation of miRNAs is associated with various malignancies, acting as either tumor suppressors or oncogenic promoters. LncRNAs, transcripts longer than 200 bp, regulate gene expression at epigenetic, transcriptional, and post-transcriptional levels through specific regulatory actions [64]. They exert diverse biological functions such as inducing chromatin remodeling, recruiting transcriptional machinery, acting as competitive endogenous RNAs for microRNAs, and modulating protein-protein interactions [64]. LncRNAs play a crucial role in regulating tumor epithelial/mesenchymal plasticity (EMP) and cancer stemness, influencing tumor initiation and progression [64]. They can increase the expression of transcription factors like Sox2 and ZEB1/2, often by sponging miRNAs, and promote the activation of signaling pathways vital for CSCs, such as β-catenin signaling [65]. LncRNAs are also important for telomere protection and elongation and can influence chemoresistance [65].

Exosomes are small vesicles of endocytic cargo that ferry bioactive molecules such as proteins, lipids, RNA, DNA, and metabolites [66]. They are crucial for inter- and intracellular communication between cancer cells, normal cells, and the environment, acting as carriers of information and contributing to reprogramming and metastasis of malignant cells [66]. Tumor-originated exosomes can induce NF-κB signaling in macrophages, increasing inflammatory factors that promote tumor survival and progression [66]. They can also block positive immune responses, allowing cancer cells to evade immune checkpoints [66]. LncRNAs, such as H19 and Sox2ot, can be packaged into exosomes and transferred between CSCs and other cell types in the TME, facilitating communication and enhancing EMT and stemness [65]. CAF-derived exosomes, for instance, mediate tumorigenesis by carrying TGF-β, CD81, or CD9, leading to ECM remodeling, and can transfer miRNAs that contribute to therapy resistance [67].

Metabolic reprogramming and adaptive plasticity

Metabolic transformation is a fundamental hallmark of cancer, providing tumor cells with the necessary energy and structural resources for rapid proliferation [68]. The key feature of cancer-related metabolism is its overall plasticity, allowing cancer cells to adapt to various conditions and resist different therapies [32]. This metabolic reprogramming is heterogeneous and dynamic, influenced by the tissue of origin, driving mutations, and the microenvironment [32].

Key features include increased aerobic glycolysis (the Warburg effect), which cooperates with oxidative phosphorylation (OXPHOS) to establish a “hybrid” metabolic state, deregulated tricarboxylic acid (TCA) cycle, altered lipid and glutamine metabolism, and increased one-carbon metabolism [32]. Cancer cells are highly opportunistic, adapting to local metabolite levels within the TME [68]. For example, highly proliferative activated T cells and cancer cells both heavily rely on glucose metabolism, leading to cellular competition for nutrients [68]. The accumulation of lactate, a glycolytic product, is detrimental to T cell effector function and anti-tumor response, and can induce a pro-tumoral M2 phenotype in tumor-associated macrophages (TAMs) [68].

Metabolic reprogramming also mediates CSC plasticity and contributes to chemoresistance. Lineage tracing revealed unprecedented cellular plasticity in glioma cells, allowing them to reprogram from a differentiated to a CSC-like state, driven by temozolomide-induced hypoxia-inducible factors (HIFs), which is a mechanism for chemoresistance acquisition [69]. Hypoxia, a common feature of solid tumors, can increase cancer cell plasticity and drug resistance through metabolic reprogramming and translational reprogramming [18]. CSCs tend to preferentially localize within hypoxic niches, developing adaptive mechanisms that lead to more aggressive behavior and protection against treatment [70]. This involves a metabolic shift towards glycolysis, reducing reactive oxygen species (ROS) levels, and different responses to varying oxygen concentrations that can induce EMT and invasion [70] (Fig. 2). This metabolic flexibility provides an actionable clinical target: inhibitors of specific metabolic pathways that CSCs preferentially rely on (such as glycolysis, glutaminolysis, or oxidative phosphorylation) could be used to selectively ‘starve’ the CSC pool, thereby compromising their survival and sensitizing them to other treatments.

Fig. 2.

Schematic illustration of metabolic reprogramming and adaptive plasticity in cancer. Tumor cells exhibit increased glucose uptake and glycolysis, producing lactate that impairs T cell function and promotes a pro-tumoral M2-like phenotype in tumor-associated macrophages (TAMs). Concurrently, cancer cells utilize oxidative phosphorylation (OXPHOS) and the TCA cycle, establishing a hybrid metabolic state. Under hypoxia, stabilization of hypoxia-inducible factors (HIFs) drives metabolic adaptation, enhances epithelial-to-mesenchymal transition (EMT), and promotes CSC-like phenotypes, thereby fueling invasion, metastasis, and therapeutic resistance

Implications of CSC plasticity in tumor initiation, progression, metastasis, and therapeutic resistance

The inherent plasticity of CSCs has profound implications across all stages of cancer development and its response to therapy.

Tumor initiation

Both intrinsic and acquired plasticity are crucial for tumor initiation. Cancer arises more commonly in tissues with greater plasticity and proliferative potential, such as the liver and colon, while tissues with little regenerative potential like the heart are largely resistant to tumorigenesis. This suggests that the cellular plasticity involved in healthy tissue regeneration contributes to tumor initiation. Progenitors and stem cells transform readily, whereas differentiated cells are resistant to transformation, due to greater transcriptional plasticity maintained by chromatin-modifying enzymes. Plasticity in tumor initiation can also be induced by specific oncogenic mutations or stressors. For instance, overexpression of SOX2 can drive lung squamous cell carcinomas from multiple lung epithelial cell types. Tissue injury or inflammation can increase plasticity and facilitate tumor initiation even in transformation-resistant cells [17–19].

Tumor progression

Plasticity is a central feature of fully formed tumors, with cancer cells exhibiting higher plasticity than terminally differentiated normal cells. This is driven by the selection of epigenetic states that permit transitions within a wide range of cellular phenotypes. While the traditional CSC model proposes a hierarchical structure, recent evidence suggests a more dynamic process where non-CSCs can de-differentiate and acquire stemness in a vacated niche, as seen in colon adenocarcinoma. A highly plastic cell state (HPCS) distinct from CSCs and adult stem cells can emerge during tumor evolution, predicted to have the capacity to acquire the largest number of distinct downstream cancer cell states and generate more phenotypic heterogeneity. This HPCS appears to be a precursor to epithelial-mesenchymal transition (EMT), which endows carcinoma cells with high invasive and metastatic potential [18, 62].

Metastasis

EMT is a key mechanism driving metastasis, enabling cancer cells to acquire mesenchymal characteristics, increasing their invasiveness and resistance to programmed cell death [27]. This process is closely linked to the generation of a CSC-like population that promotes tumor recurrence and spread [1].Cells in a hybrid EMT state, exhibiting mixed epithelial and mesenchymal features, are often more competent at forming metastases and possess increased plasticity [17]. For example, in pancreatic cancer, the transcription factor ZEB1 promotes EMT and stemness by repressing the miR-200 family and interacting with CD44 isoforms, creating a self-sustaining loop that increases tumorsphere initiation and metastasis [27]. Nestin, another CSC marker, also regulates EMT and promotes metastasis in pancreatic cancer [27].

Therapeutic resistance

Cancer cell plasticity significantly contributes to treatment resistance by promoting adaptation to therapeutic pressure through differentiation to treatment-resistant states. This is a persistent challenge in oncology, often undermining treatment efficacy and implicated in approximately 90% of cancer mortalities [17].

Reactivation of developmental programs, particularly EMT pathways, is integral to drug resistance. Heightened expression of mesenchymal and stromal markers correlates with pronounced resistance to various treatments, including chemotherapy, targeted therapy, radiation, and immunotherapy [17]. Cancer cells can strategically shift between a drug-sensitive, proliferative state and a drug-tolerant, slow-cycling “persister” state, undergoing transient, reversible changes to endure lethal drug concentrations. This “emergence of altered cellular identity” involves extensive reprogramming, associated with epigenetic and transcriptional modifications [17]. For instance, lung adenocarcinoma treated with EGFR TKIs can lose dependence on EGFR signaling by transforming into small cell lung cancer or squamous cell carcinoma [71]. Prostate adenocarcinomas develop resistance to antiandrogens by shifting phenotype to neuroendocrine prostate cancer, a process significantly enhanced by loss of TP53 and RB1 [18].

Furthermore, CSCs contribute to drug resistance through various mechanisms: they often exist in a quiescent or dormant state, protecting them from therapies targeting rapidly dividing cells [15]. They can express drug efflux pumps (e.g., ATP-binding cassette family transporters) that remove chemotherapeutic agents from the cell [42]. CSCs also possess enhanced DNA repair capabilities, allowing them to survive DNA damage induced by chemotherapy [42]. The p38 MAPK pathway, for example, is involved in drug efflux, ALDH activity, and promoting quiescence and defective DNA repair in CSCs, all contributing to chemoresistance. The reprogramming of chemotherapy-induced senescent cells into a stem-like state also poses a significant hindrance, as these cells can acquire high tumor initiating ability and upregulate canonical Wnt signaling, leading to cancer relapse [72].

The tumor microenvironment: a critical regulator of CSC fate

Cellular components of the TME and their interplay with CSCs

The cellular constituents of the TME are not passive bystanders but active participants in shaping CSC fate, providing essential cues for their maintenance, proliferation, and adaptive responses.

Cancer-associated fibroblasts (CAFs)

Cancer-associated fibroblasts (CAFs) are a major, heterogeneous stromal component activated by cancer cells (e.g., via TGF-β and IL-6) to become key drivers of progression [41, 73]. Their primary roles in supporting the CSC niche involve extensive ECM remodeling and the paracrine secretion of numerous growth factors, cytokines, and proteases [67, 73].

CAFs contribute to tumor progression through extensive ECM remodeling. They synthesize and secrete various ECM components, including collagens, elastin, proteoglycans, and glycoproteins, and also secrete enzymes like matrix metalloproteinases (MMPs) that degrade the ECM [67]. This dynamic remodeling leads to increased ECM stiffness, a major indicator of cancer’s progressive and invasive potential, which in turn promotes cancer cell proliferation and resistance [67]. CAFs can also create “invasion-friendly pathways” or “tracks” for invading cancer cells, facilitating their migration [67].

Beyond structural support, CAFs secrete numerous soluble factors that act in a paracrine manner to influence tumor growth, metastasis, and immune modulation. For instance, CAF-secreted TGF-β contributes to tumor epithelial-to-mesenchymal transition (EMT) and invasiveness [67]. IL-6 from CAFs promotes tumor angiogenesis by inducing VEGF release and is associated with an immunosuppressive stroma in pancreatic cancer [67]. Hepatocyte growth factor (HGF) secreted by CAFs supports CSC properties by inducing Wnt/β-catenin signaling, even restoring stem-like features in non-CSCs and expanding the CSC poo [73]l. In lung cancer, CAFs induce the NANOG transcription network through paracrine IGF-II/IGF-1R signaling [73]. In prostate cancer, CAFs elicit EMT and increase stemness by secreting MMPs [73].

CAFs significantly impact the immune microenvironment. They recruit macrophages to the stroma and play a role in differentiating them into the pro-tumor M2 phenotype [67]. They can suppress Natural Killer (NK) cell function and recruit neutrophils [67]. CAFs also directly contribute to T-cell suppression and deactivation by secreting immunosuppressive factors like TGF-β, IL-10, and arginase I, and can exclude T-cells from the tumor through ECM remodeling [67]. This intricate crosstalk between CAFs and tumor cells modulates the cell cycle, stemness, EMT, and therapeutic resistance, as seen in Head and Neck Squamous Cell Carcinomas (HNSCC). For example, CAF-secreted IL-6 can induce higher stemness and EMT in HNSCC cells, partially via the MAPK pathway, and Wnt activation by CAFs increases CSC characteristics [41].

Tumor-associated macrophages (TAMs) and other immune cells

With This Trimmed Version: Tumor-associated macrophages (TAMs) are an abundant, highly plastic immune population in the TME, primarily derived from bone marrow monocytes recruited by tumor signals. They are pivotal in tumorigenesis and are broadly categorized into anti-tumor (M1) and pro-tumor (M2) phenotypes [74].

TAMs can be broadly categorized into pro-inflammatory M1 type (anti-tumor) and anti-inflammatory M2 type (pro-tumor). While M1 macrophages induce a Th1 immune response and kill tumor cells, TAMs in most tumors are predominantly of the M2 phenotype, which promotes tumor infiltration, metastasis, and creates a favorable microenvironment for tumor survival and angiogenesis [74].

TAMs promote tumor proliferation, invasion, and metastasis through various pathways. They activate NF-κB and STAT3 by expressing inflammatory factors like TNF-α, IL-6, and IL-11. M2 TAMs secrete epidermal growth factor (EGF), which enhances tumor cell invasion, and matrix metalloproteinases (MMPs) and plasmin, linked to tumor aggressiveness. They are strongly linked to angiogenesis, regulating it through factors like VEGF, FGF1, PDGF, HGF, and MMPs [74]. TAMs also create an immunosuppressive TME by producing immunosuppressive chemokines and factors like IL-10 and TGF-β, which suppress T cell function and promote the expansion of regulatory T cells (Tregs) [74].

The crosstalk between CSCs and TAMs is critical and bidirectional. CSCs contribute to macrophage infiltration and polarization through secreted chemokines like CCL2, CCL3, and periostin [75]. TAMs, in turn, reciprocally support CSC stemness and the CSC niche through various secreted factors and direct interactions [75]. Direct ligand/receptor interactions, such as hyaluronic acid/CD44, CD90/CD11b, and EPHA4/Ephrin binding between CSCs and TAMs, activate signaling pathways (e.g., PI3K-EIF4EBP1-SOX2, SRC, NF-kappa-B) that maintain CSC stemness [76]. Indirectly, TAMs secrete cytokines like IL-6, TGF-β1, TNF-α, CCL18, and CCL2, which induce CSC proliferation, EMT, and stemness (e.g., via STAT3, WNT/β-catenin pathways) [75]. Exosomes released by TAMs can also regulate CSC behavior by enhancing stemness and chemoresistance, and conversely, CSC-derived exosomes can polarize TAMs toward a tumor-promoting M2 phenotype [76]. In colorectal cancer, CSCs can suppress T cell proliferation by secreting high levels of IL-4 and promote PD-L1 expression via Wnt/β-catenin signaling, leading to immune resistance [77]. They also remodel the immunosuppressive microenvironment by affecting TAMs and secreting miR-146a-loaded exosomes that reduce CD8 + T cell infiltration [77]. This CSC-driven manipulation of the immune environment is a key survival strategy. For example, in colorectal cancer, CSCs can suppress T cell proliferation by secreting high levels of IL-4 and can promote PD-L1 expression via Wnt/β-catenin signaling, leading to direct immune resistance. They further remodel the immunosuppressive microenvironment by affecting TAMs and secreting miR-146a-loaded exosomes that reduce CD8 + T cell infiltration [77].

Other immune cells within the TME, including T cells, Natural Killer (NK) cells, Myeloid-Derived Suppressor Cells (MDSCs), and Dendritic Cells (DCs), also play crucial roles. Cytotoxic CD8 + T cells provide natural defense by killing tumor cells, but their function is often impaired in the TME due to competition for nutrients like glucose and amino acids, and accumulation of lactate [68]. MDSCs suppress immune responses and can enhance cancer cell stemness through factors like miR-101, GM-CSF, piRNA-823, IL-6, nitric oxide, and exosomal S100A9 [75]. DCs are essential for antigen presentation and T cell priming, but their function can be limited by tumor-induced metabolic reprogramming [68]. The mutual plasticity between cancer and immune cells means that immune cells can induce tumor cell plasticity (e.g., cytotoxic cells inducing EMT, myeloid cells promoting invasive programs), while cancer cells actively manipulate immune cell plasticity (e.g., altering T cell differentiation, NK cell dynamics, and myeloid reprogramming) to promote tumorigenic growth and immune evasion [43].

Endothelial cells and pericytes

Endothelial cells (ECs) are crucial components of the tumor vasculature, which is often abnormal and leaky, leading to hypoxia and hindering drug delivery [78]. ECs form a “vascular niche” for CSCs, providing microenvironmental signals essential for CSC homeostasis [36]. This interaction is bidirectional: CSCs secrete vascular endothelial growth factor (VEGF) to induce local angiogenesis, while ECs produce nitric oxide, which activates Notch signaling in glioma CSCs [72]. Further complicating this niche signaling, recent work by Zhou et al. in glioblastoma identified that ECs also secrete the Wnt inhibitor NOTUM. Paradoxically, this suppression of the Wnt/β-catenin pathway was found to be essential for maintaining glioblastoma stemness and promoting radioresistance, highlighting the complex and context-dependent signals CSCs receive from their vascular niche [79]. ECs can also secrete basic fibroblast growth factor (bFGF) to enhance the expression of stem-cell markers and sphere-forming ability of differentiated glioblastoma cells [72]. The vascular niche is critical in inducing quiescence in circulating tumor cells (CTCs) and promoting the survival of CSCs and CTCs during metastasis, contributing to their dormancy [72]. Pericytes, which support blood vessel function and stability, are also involved in neovascularization and basement membrane remodeling during tumorigenesis [80]. Glioblastoma stem cells can generate vascular pericytes to support vessel function and tumor growth, suggesting their role in remodeling the CSC niche [80]. In premalignant oral epithelium, TAMs and vascular ECs nurture CSC development and support stemness through a reciprocal exchange of cytokine mediators, creating a positive feedback loop [81]. This reciprocal support is a common feature; in some contexts, such as premalignant oral epithelium, TAMs and vascular ECs have been shown to nurture CSC development through a reciprocal exchange of cytokine mediators, creating a positive feedback loop that sustains malignant transformation [81].

Mesenchymal stem cells (MSCs)

Mesenchymal stem cells (MSCs) are multipotent stromal cells that reside in various tissues and can be recruited to the TME, where they create an advantageous environment for CSC restoration [73]. MSCs support CSC properties and self-renewal by secreting various paracrine factors, including CXCL12, IL-6, and IL-8, which promote CSC self-renewal and proliferation [73]. They also secrete BMP antagonists that help maintain the undifferentiated state of CSCs [80]. MSCs can induce EMT and increase the stemness properties of cancer cells through the secretion of MMPs, as seen in prostate cancer, or SDF-1 and TGF-β1 in breast cancer [73]. MSCs can also support CSC development through cell fusion, forming hybrid tumor cells, and by releasing soluble molecules (metabolites, hormones) and exchanging microvesicles and exosomes [82]. Their migration to tumor sites is similar to their migration to injured or ischemic sites, attracted by chemoattractant factors like CXCL16, SDF-1, CCL-25, and IL-6 secreted by tumor cells and tumor-associated immune cells [83] (Fig. 3).

Fig. 3.

Schematic representation of the cellular components of the tumor microenvironment (TME) and their crosstalk with cancer stem cells (CSCs). Cancer-associated fibroblasts (CAFs) regulate CSCs via IL-6, HGF, MMPs, and TGF-β. Tumor-associated macrophages (TAMs) and other immune cells (e.g., T cells, DCs, MDSCs) influence CSC stemness through cytokines and immunomodulatory factors such as IL-10, TGF-β, and EGF. Endothelial cells and pericytes provide vascular niche support via VEGF and bFGF. Mesenchymal stem cells (MSCs) sustain CSCs by secreting IL-6, IL-8, CXCL12, and SDF-1. This complex bidirectional interplay collectively maintains CSC properties, enhances EMT, promotes therapeutic resistance, and drives tumor progression

Non-cellular components of the TME and their influence on CSCs

Beyond cellular interactions, the non-cellular components of the TME profoundly influence CSC behavior, plasticity, and therapeutic response.

Extracellular matrix (ECM) and its dynamic remodeling

The extracellular matrix (ECM) is a crucial component of the local microenvironment, described as a complex network of macromolecules with distinct physical, biochemical, and biomechanical properties [84]. While tightly regulated in normal tissues, the ECM is commonly deregulated and disorganized in cancer, directly promoting cellular transformation and metastasis [85].

ECM anomalies deregulate stromal cell behavior, facilitate tumor-associated angiogenesis and inflammation, and thus lead to the generation of a tumorigenic microenvironment [85]. The ECM is constantly remodeled by cells, which degrade and reassemble it, playing an active role in sculpting their surrounding environment and directing their own phenotypes [86]. This dynamic reciprocal communication between cells and the ECM is fundamental for tissue development, homeostasis, and wound healing [86].

In cancer, ECM remodeling in CSC niches facilitates abnormal interactions between CSCs and the ECM, which is critical for regulating phenotypic plasticity, senescence, and quiescence in CSCs [72]. ECM components initiate cellular responses by binding to cell surface receptors (e.g., integrins), sequestering growth factors and chemokines, and providing tensile strength cues [72]. For instance, periostin, an ECM protein from cancer-associated fibroblasts (CAFs), augments Wnt signaling in CSCs, promoting their stemness and growth [72]. Tenascin C, another ECM protein, enhances Notch and Wnt pathways, promoting self-renewability and viability of CSCs and re-initiating metastatic outgrowth [72]. Hyaluronan, a major ECM component often overexpressed in aggressive cancers, induces EMT transcription factors like SNAIL and TWIST through TGF-β signaling, leading to CSC population expansion and promoting cell plasticity [72]. The integrin-FAK pathway, an ECM signaling pathway, is also linked to CSC functions [72]. Furthermore, senescent cells remodel the ECM and express matrix metalloproteases (MMPs) as senescence-associated secretory phenotype (SASP) factors, creating an environment conducive to invasion and metastasis [72].

Soluble factors

The TME is a rich milieu of soluble factors, including various growth factors, cytokines, chemokines, RNA, DNA, and metabolites, which orchestrate numerous ways to support cancer survival and progression [66].

Growth factors such as Vascular Endothelial Growth Factor (VEGF), Hepatocyte Growth Factor (HGF), Fibulin-3, Progranulin, Insulin-like Growth Factor-1 (IGF-1), Platelet-Derived Growth Factor (PDGF), and Sphingosine-1-phosphate play major roles in angiogenesis, EMT, and invasion [66]. The clinical relevance of targeting these factors is high; for instance, Chen et al. recently developed a soluble dual-functional nanobody that simultaneously blocks VEGF and HGF, which successfully inhibited tumor growth and angiogenesis in preclinical models [87]. Cytokines like Interleukin-33 (IL-33), IL-23, IL-10, IL-4, IL-6, Transforming growth factor–β (TGF-β), Chemokine CCL5, and CXC subfamily chemokines (IL-8, GROα, IP-10, MIG) are involved in inflammation, immune suppression, tumor progression, and DNA damage protection [66]. IL-6, for example, is a major component of almost all TMEs, activating inflammation and controlling pro-cancer activities like progression, malignancy, and anti-death signaling pathways [66]. TGF-β is essential for self-renewal, differentiation, maintenance, and survival of CSCs, leading to tumor progression and metastasis [66].

Exosomes, small vesicles containing proteins, lipids, RNA, DNA, and metabolites, are important for reprogramming and metastasis of malignant cells [66]. They mediate inter- and intracellular communication, transport materials to cancer cells, block immune responses, and contribute to drug resistance by removing drugs and promoting factors that negatively regulate immune responses. Extracellular RNA (e.g., microRNAs) and DNA also contribute to TME-based heterogeneity and communication. Metabolites, such as lactate, ketone bodies, fatty acids, glutamine, and other amino acids, secreted by cancer-associated fibroblasts (CAFs), can act as energy-rich fuels for cancer cells, influencing phenotypic diversity. An altered, often decreased, extracellular pH is considered a hallmark of cancer that promotes tumor progression by damaging ECM proteins, leading to invasion and metastasis [66].

Hypoxia and pH gradients

Hypoxia, or low oxygen conditions, is a hallmark of the TME and plays a crucial role in promoting cancer stem cell (CSC) resistance and adaptive mechanisms. CSCs tend to preferentially localize within hypoxic niches, which helps them develop adaptive mechanisms mediated by modified responses to various stressors, leading to more aggressive behavior [70].

Hypoxia-inducible factors (HIFs), particularly HIF-1α and HIF-2α, are key regulators of CSC phenotypes. HIF-2α is specifically responsible for cell survival and stemness activation in glioma CSCs and promotes a stem-like phenotype in non-stem tumor cell populations by upregulating stemness-associated transcriptional factors such as NANOG, OCT4, and c-MYC [72]. Hypoxia also stimulates epithelial-mesenchymal transition (EMT) during tumor progression and metastasis, indicating its role in regulating cell plasticity in CSCs [72].Furthermore, hypoxia induces quiescence in tumor cells and CSCs through HIF-1α, Akt, and CREBBP/Creb Binding Protein (CBP), and inhibits senescence by downregulating p21 via HIF-1α [72].

The hypoxic environment is interconnected with other TME components. CSC resistance in hypoxic regions is maintained through intense interactions with immune cells (Tregs, TAMs, MDSCs) and CAFs, which form supportive niches that help maintain the quiescent state in CSCs and guide cellular invasion [70]. Hypoxia also causes a metabolic shift in CSCs towards glycolysis, which helps reduce accelerated reactive oxygen species (ROS) levels through glutathione (GSH) induction [70]. The acidic pH, often a consequence of increased glycolysis in hypoxic conditions, further contributes to tumor progression by damaging ECM proteins and promoting invasion and metastasis [66] (Fig. 4).

Fig. 4.

Schematic illustration of non-cellular components of the tumor microenvironment (TME) and their influence on cancer stem cells (CSCs). The extracellular matrix (ECM) undergoes dynamic remodeling through proteins such as tenascin and hyaluronan, which signal via integrins to promote stemness, EMT, and metastasis. Soluble factors—including growth factors (VEGF, HGF), cytokines (IL-6, IL-10, TGF-β), chemokines, exosomes, nucleic acids, and metabolites—enhance CSC survival and plasticity. Hypoxia stabilizes hypoxia-inducible factors (HIFs), leading to acidic pH and activation of pathways governing CSC quiescence, self-renewal, metabolic adaptation, EMT, and invasion

Therapeutic strategies targeting CSC plasticity and TME crosstalk

Given the profound roles of CSC plasticity and TME crosstalk in driving tumor progression and therapeutic resistance, developing effective anti-cancer strategies necessitates a multi-pronged approach. Current research focuses on directly targeting CSC-intrinsic plasticity mechanisms, modulating the TME to impair CSCs, and exploring synergistic combined therapeutic approaches.

Directly targeting CSC-intrinsic plasticity

Targeting the inherent adaptability of CSCs is a crucial strategy to prevent tumor recurrence and metastasis. This involves inhibiting key signaling pathways, modulating epigenetic mechanisms, and targeting specific RNA modifications.

Small molecule inhibitors of key signaling pathways

The aberrant activation of various signaling pathways in CSCs provides numerous targets for small molecule inhibitors.

• Wnt Inhibitors: The Wnt signaling pathway is frequently deregulated in cancers and is critical for CSC self-renewal and maintenance [88]. Beyond canonical drivers, TMEM64 was recently shown to aggravate glioma malignancy via Wnt/β-catenin activation, underscoring membrane regulators as actionable Wnt nodes [89]. Similarly, Yan et al. identified that disrupting TFRC-mediated iron metabolism leads to the degradation of Wnt pathway components, offering another innovative strategy to inhibit Wnt signaling in colorectal cancer [90]. Inhibitors targeting different components of this pathway have entered clinical trials. Inhibitors targeting different components of this pathway have entered clinical trials. For example, Ipafricept (OMP-54f28/FZD8-Fc), a recombinant fusion protein antagonizing Wnt signaling, is being investigated in combination with chemotherapy for pancreatic and ovarian cancers [42]. PRI-724, a β-catenin inhibitor, has completed Phase I studies for advanced myeloid malignancies and pancreatic cancer [42]. CWP232291, another β-catenin activity inhibitor, is in Phase I/II trials for AML and myeloma [42]. Other Wnt inhibitors like LGK974 and ETC-159 (Porcupine inhibitors) and OMP-18R5 (Fzd receptor antibody) are also in clinical development [45]. Clinically, these agents are largely in Phase I/II trials, often limited by on-target toxicities in Wnt-dependent healthy tissues (e.g., bone and gut) and the rapid emergence of resistance via pathway reactivation.

• Notch Inhibitors: The Notch pathway plays a core role in malignant tumor development by regulating cell differentiation, proliferation, and apoptosis [91]. Gamma-secretase inhibitors (GSIs), which block Notch activation, were among the first Notch inhibitors in clinical studies. MK-0752 and RO4929097 are GSIs that have been tested in various cancers, showing modest clinical benefits, often in combination therapies [42]. Nirogacestat (PF-03084014) is in Phase II trials for desmoid tumors with promising activity [45]. DLL4-targeting antibodies like Demcizumab and Enoticumab are also in clinical development for solid tumors [42]. Maturity is advancing (e.g., Nirogacestat for desmoid tumors), but broad use of GSIs is hampered by severe on-target gastrointestinal toxicities, as Notch signaling is essential for gut homeostasis.

• Hedgehog (Hh) Inhibitors: The Hh pathway is aberrantly activated in many cancers, promoting proliferation and survival [58]. Smoothened (Smo) inhibitors are the most successful Hh-targeting drugs. Vismodegib and Sonidegib are FDA-approved for basal cell carcinoma and are in trials for other cancers like medulloblastoma, breast, and pancreatic cancer [42]. Glasdegib is approved for AML [42]. Novel inhibitors of terminal components like arsenic trioxide (ATO) and GANT-61 are also being developed [45]. While several SMO inhibitors are FDA-approved (e.g., for basal cell carcinoma), their utility is often limited by frequent on-target toxicities (e.g., muscle cramps, alopecia) and acquired resistance, commonly driven by mutations in the SMO binding site.

• JAK/STAT Inhibitors: The STAT3 pathway regulates CSC self-renewal, differentiation, and apoptosis [45]. Napabucasin (BBI608), a first-in-class STAT3 inhibitor, is in Phase III clinical trials for metastatic colorectal carcinoma and pancreatic cancer, showing strong anti-CSC effects [45]. These inhibitors, including Napabucasin which has reached Phase III trials, show promise but face challenges of dose-limiting toxicities (e.g., myelosuppression for broader JAK inhibitors) and resistance through activation of parallel survival pathways.

• PI3K/Akt/mTOR Inhibitors: This pathway is crucial for cell growth, proliferation, and is associated with drug resistance [54]. GDC-0084 is in a Phase II trial for glioblastoma multiforme (GBM) [45]. Other PI3K inhibitors like Alpelisib, Buparlisib, and SF1126 are in various clinical phases [45]. Despite several approvals (e.g., Alpelisib), this class is frequently associated with significant on-target metabolic toxicities (e.g., hyperglycemia, rash), and resistance often emerges through compensatory activation of the parallel MAPK pathway.

• NF-κB Inhibitors: NF-κB is constitutively activated in several tumors and linked to CSC self-renewal, expansion, and chemoresistance [47]. Small-molecule inhibitors like parthenolide and pyrrolidinedithiocarbamate preferentially target breast cancer stem cells [48]. Triptolide has shown to reverse EMT and reduce self-renewal in pancreatic cancer [48]. Direct targeting remains clinically challenging, as most inhibitors are in early-phase or preclinical development due to significant off-tumor toxicities related to NF-κB’s essential role in normal immune cell function.

• MAPK Inhibitors: The p38 MAPK pathway is involved in CSC development, maintenance, metastasis, and chemoresistance [56]. Inhibition of p38 can decrease stemness factors, compromise metastasis, and revert mesenchymal-like tumor cells to a chemosensitive epithelial state [56, 92, 93]. While p38 inhibitors have been explored, their clinical development in oncology has been slow, often due to a narrow therapeutic window and compensatory signaling, though targeting the ERK pathway (e.g., MEK/BRAF inhibitors) is a clinical standard in relevant cancers.

Epigenetic modulators

Epigenetic modifications are key drivers of CSC plasticity, and targeting these reversible changes offers a promising therapeutic avenue.

• Targeting Unique Gene Expression in CSCs: Since CSCs have unique gene expression profiles necessary for self-perpetuation, such as chromatin remodelers, these genes can serve as drug targets [34]. For example, silencing Bmi1, a gene essential for self-replication, prevents tumor formation in glioblastoma [34].

• LSD1 Inhibition: LSD1 (lysine-specific demethylase 1) inhibition induces differentiation and cell death in Merkel cell carcinoma and potentiates responsiveness to retinoic acid in myeloid malignancies [21].

• Other Epigenetic Drugs: HDAC inhibitors, DNMT inhibitors, EZH2 inhibitors, and BET inhibitors are in clinical studies for hematologic malignancies and show promise in modulating EMT and tumor plasticity, potentially increasing sensitivity to conventional therapies and promoting a more permissive immune environment [51]. This class includes several FDA-approved agents (e.g., for hematologic malignancies), but their use is often constrained by broad, systemic toxicities (particularly hematological) and resistance due to the functional redundancy of epigenetic regulators.

• KDM1B Inhibition: Lysine-specific histone demethylase 1B (KDM1B) promotes IFN-I-induced transcriptional rewiring of cancer cells towards stemness and immune escape [94]. Inhibiting KDM1B offers an attractive therapeutic approach to increase the beneficial effects of anti-tumoral therapies by preventing CSC expansion and boosting tumor resistance [94].

Targeting m6A RNA modification (writers, erasers, readers)

N6-methyladenosine (m6A) modification is the most prevalent internal modification in eukaryotic mRNAs and plays a critical role in gene expression regulation [95]. Aberrant m6A modification is closely associated with cancer occurrence, development, progression, and prognosis, and dysregulated m6A regulators have been identified as novel anticancer drug targets [96]. Dysregulation of m6A modification is strongly implicated in modulating tumor stemness and CSC maintenance [96]. This regulatory control often involves complex interactions with non-coding RNAs. For instance, in glioma, recent work by Wu et al. identified a specific m6A-lncRNA axis, showing that m6A-mediated upregulation of the lncRNA CHASERR promotes glioma progression by modulating a downstream miR-6893-3p/TRIM14 axis [97]. The profound, multifaceted role of m6A machinery (writers, erasers, and readers) in governing gene expression is now recognized as a “new frontier” for therapeutic development, not only in cancer but also in other complex diseases like epilepsy, underscoring the fundamental nature of this epitranscriptomic system [79].

Pharmacological inhibition of the dysregulated m6A machinery exhibits robust anti-tumor efficacy [98] (Tables 2 and 3). This field is in its infancy, with most agents in preclinical or early Phase I development; consequently, the long-term toxicity profiles and mechanisms of acquired resistance are not yet well understood.

Table 2.

Key m6A RNA modifiers and their roles in cancer stemness and TME crosstalk

m6A Regulator Type	Protein	Role in m6A Modification	Role in Cancer Stemness and TME Crosstalk	References
Writers	METTL3	Catalyzes m6A deposition on mRNA, often with METTL14.	Functions as an oncogene, promotes cancer cell growth and tumorigenicity in various cancers (e.g., AML, prostate cancer, lung adenocarcinoma). Inhibition (e.g., STM2457) shows anti-leukemic effects, reduces leukemia stem cells, and suppresses prostate cancer progression. Promotes stemness in lung adenocarcinoma by stabilizing NFE2L3 mRNA.	[99, 100]
	METTL14	Essential cofactor for METTL3 in m6A deposition. 75	Knockdown promotes human glioblastoma stem cell (GSC) growth, self-renewal, and tumorigenesis	[99]
Erasers	FTO	First identified m6A demethylase. 75	Acts as an oncogene in various solid tumors (e.g., AML, glioblastoma, breast, pancreatic cancer). FTO inhibitors (e.g., CS1, CS2) suppress tumor progression, attenuate cancer cell metabolism, improve drug response, overcome tumor immune evasion, and reduce leukemia stem/initiating cell (LSC/LIC) self-renewal. FTO inhibition enhances radiotherapy efficacy in head and neck cancer.	[99]
	ALKBH5	Another m6A demethylase. 75	Context-dependent role (tumor suppressor/carcinogen). Promotes breast cancer stemness, doxorubicin resistance, and HER2-targeted therapy resistance. Highly expressed in glioblastoma stem-like cells, maintaining FOXM1 expression. Downregulated in gastric and bladder cancers, where it inhibits tumor growth/metastasis and sensitizes to cisplatin.	[101]
Readers	YTHDF2	Recognizes m6A-containing RNA to mediate downstream effects. 75	Contributes to cancer progression. Knockdown suppresses cervical cancer stemness and inhibits proliferation by sustaining GLI2 stability. Promotes non-small cell lung cancer glycolysis and stemness. Maintains oncogene expression (e.g., MYC, IGFBP3) in glioblastoma stem cells. Inhibition potentiates radiotherapy and overcomes MDSC-induced immunosuppression.	[102, 103]
	YTHDC1	m6A modification reading protein.	Upregulated in head and neck squamous cell carcinoma (HNSCC) spheres and tumor tissues. Positively correlates with malignant epithelial cell stemness capacity and is involved in regulating stemness maintenance in HNSCC.	[104]

Table 3.

Selected clinical trials for m6A pathway inhibitors in cancer

NCT ID	Investigational Drug	Target Cancer Type(s)	Clinical Phase	Sponsor	Status	Ref.
NCT02973789	STC-15 (METTL3 Inhibitor)	Advanced Malignancies	Phase 1	STORM Therapeutics	Completed	[105]
NCT06762925	METTL3 Peptide Inhibitors	Urinary Tract Tumors (Bladder, Prostate, Kidney)	Observational (Study Start Est. 2025)	Xijing Hospital	Not yet recruiting
NCT00651261	Midostaurin (FLT3 inhibitor, affects m6A indirectly)	Newly Diagnosed FLT3 Mutated Acute Myeloid Leukemia (AML)	Phase 3	Alliance for Clinical Trials in Oncology	Completed	[106]

While some clinical trials for general targeted therapies (e.g., TAPUR study, NCT02693535) or specific ADCs (DB-1303/BNT323, NCT05150691) were found, their explicit link to m6A inhibition as a primary mechanism was not clearly stated in the provided research material. Midostaurin is a FLT3 inhibitor, but its connection to m6A is indirect as “abnormal N6-methyladenosine (m6A) modification is closely associated with the occurrence, development, progression and prognosis of cancer, and aberrant m6A regulators have been identified as novel anticancer drug targets”. Therefore, only trials with direct or strongly implied m6A targeting are included

Strategies to induce CSC differentiation or eliminate plasticity

Beyond direct inhibition of pathways, strategies to induce CSC differentiation or eliminate their plasticity are being explored. This involves targeting unique gene expression in CSCs, such as Bmi1, essential for self-replication in glioblastoma [34].Blocking differentiation and transdifferentiation, for instance, preventing GBM cells from transdifferentiating into endothelial cells that form tumor blood vessels, is another avenue [34]. Identifying specific druggable targets, including kinases and transcription factors unique to CSCs, is crucial for their elimination [34].

All-trans retinoic acid can promote leukemic cell differentiation in acute promyelocytic leukemia, leading to high cure rates when combined with other therapies [21]. LSD1 inhibition also induces differentiation and cell death in Merkel cell carcinoma [21]. Targeting drug-tolerant cells, which often activate specific signaling pathways for survival, can be achieved by incorporating epigenetic modulators into existing therapies or by inhibiting these activated pathways [21]. For example, disrupting the repressed chromatin state that maintains resistance to EGFR TKIs in non-small cell lung cancer by HDAC inhibition or IGF-1 receptor inhibition has shown to be lethal to drug-tolerant persister (DTP) cells.

in vitro [21]. Strategies that inhibit CSC self-renewing capacities or promote their differentiation are being explored to achieve CSC exhaustion and tumor regression [21].Targeting plasticity to prevent phenotype switching, rather than solely targeting the stable CSC state, is considered an important therapeutic approach to limit tumor recurrence [51]. This includes eliminating cells that have undergone EMT or actively reversing cellular plasticity, given its epigenetic governance.6.

Modulating the tumor microenvironment to impair CSCs

Modulating the TME is an indirect yet powerful strategy to impair CSC functions, as the TME provides essential signals for CSC maintenance, regulation of self-renewal, and homeostatic processes.

Targeting CAFs (depletion, reprogramming, metabolic interference)

Cancer-associated fibroblasts (CAFs) are critical TME components that promote tumor progression, ECM remodeling, inflammation, chemoresistance, and immunosuppression [107]. Targeting CAFs can disrupt their supportive role for CSCs. These strategies are clinically immature (Phase I/II) and face a major hurdle: the lack of CAF-specific targets, leading to on-target, off-tumor toxicities (e.g., FAP is expressed on other tissues) and the risk of depleting tumor-restraining CAF subtypes.

• Depletion of CAFs: Strategies to eliminate tumor-promoting CAFs include targeting Fibroblast Activation Protein (FAP) with FAP-targeted CAR-T cells or DNA vaccines, which can effectively delay tumor proliferation and metastasis by killing CAFs via CD8 + T cells [108]. Depletion of α-SMA (+) CAFs has also been shown to inhibit metastasis and angiogenesis [52].

• Blockage of CAF Activation: Inhibiting factors crucial for CAF activation, such as TGF-β, can enhance anti-tumor immunity [52]. Small-molecule inhibitors like Galunisertib (LY21577299) have shown anti-tumor effects in clinical trials for pancreatic and hepatocellular carcinomas [52]. Resveratrol-loaded liposomes and hydroxychloroquine (HCQ) can also block CAF activation [52]. Chemokine signaling blockade, like targeting CXCR4 with AMD3100, prevents CAF activation and promotes T-cell accumulation [52].

• Suppression of CAF-induced ECM Remodeling: Altering ECM stiffness by targeting CAFs can inhibit the recruitment of immunosuppressive cells [52]. Strategies include targeting ECM proteins like Tenascin C (TNC) with specific antibodies, depleting Hyaluronan (HA) with enzymes like PEGPH20, or inhibiting Matrix Metalloproteinases (MMPs) [52]. FAK inhibitors can also improve CAF-induced stromal stiffness [52].

• Metabolic Interference: Tumor cells can hijack CAF metabolism for energy from glutamine, lipids, and glucose, creating a metabolic coupling that fuels tumor growth [52]. Targeting this metabolic coupling, for instance, by inhibiting bromodomain and extraterminal (BET) protein to interfere with lactate-dependent lipid metabolism or by inhibiting glutamine uptake, can hinder cancer cell proliferation and restore sensitivity to therapies [52].

Targeting TAMs (depletion, reprogramming, macrophage-based cell therapies)

Tumor-associated macrophages (TAMs) are highly plastic and heterogeneous immune cells that promote tumor progression, invasion, metastasis, immunosuppression, and chemotherapy resistance [74]. These approaches are in various clinical phases, but challenges include incomplete TAM depletion, compensatory infiltration of other immunosuppressive myeloid cells, and toxicities such as hepatotoxicity (with CSF-1R inhibitors) or cytokine release syndrome (with CD40 agonists).

• Depletion of TAMs: Targeting TAMs can be achieved by blocking colony-stimulating factor-1 receptor (CSF-1R) signaling, which is essential for macrophage differentiation and survival [109]. Small molecule tyrosine kinase antagonists (e.g., PLX3397) and antibodies (e.g., AMG820, IMC-CS4, RG7155) are in clinical testing for TAM depletion [109].

• Reprogramming of TAMs: Repolarizing or re-educating macrophages from a pro-tumor M2 phenotype to an anti-tumor M1 phenotype is a promising strategy [109]. This can be achieved by CSF-1R inhibitors like BLZ945, agonistic CD40 antibodies, or low-dose radiation, which promotes an iNOS + M1 phenotype and enhances T cell recruitment [110]. Chemokine receptor blockade, such as CXCR4 inhibition with Plerixafor, can also repolarize TAMs [110].

• Macrophage-based Cell Therapeutic Technology: A new approach involves equipping macrophages with CAR (Chimeric Antigen Receptor) molecules, which is expected to overcome barriers to solid tumor treatment [111].

Vascular normalization and anti-angiogenic approaches

The abnormal and leaky vasculature within tumors leads to hypoxia and limits drug and immune cell delivery. Vascular normalization aims to reverse these abnormalities, stimulating anti-tumor immunity and enhancing therapy efficacy.

• Vascular Normalization Hypothesis: Judicious use of antiangiogenic agents, originally designed to starve tumors, can transiently normalize tumor vasculature, alleviate hypoxia, and increase delivery of drugs and anti-tumor immune cells [112]. This process prunes some abnormal vessels and remodels the remaining ones, leading to decreased leakiness and reduced hypoxia. Clinical trials with agents like cediranib have supported this concept, showing improved patient survival correlated with increased tumor blood perfusion [112, 113].

• Anti-angiogenic Therapies: VEGF-A inhibitors like bevacizumab and multi-target receptor tyrosine kinase (RTK) inhibitors like sorafenib and sunitinib are examples of agents that target tumor vasculature [110]. However, CSCs can develop resistance to these therapies, leading to increased HIF expression and tumor propagation, suggesting the need for combination strategies [82]. While many anti-angiogenic agents are FDA-approved, their efficacy is often transient, limited by hallmark toxicities (e.g., hypertension, bleeding) and rapid acquired resistance commonly driven by upregulation of alternative pro-angiogenic pathways (e.g., FGF).

• Targeting Notch Signaling in Endothelial Cells: The Notch pathway is central to controlling cell fate during angiogenesis [114]. Interfering with tumor endothelial cell growth and survival through Notch inhibition can disrupt the formation of functional tumor blood vessels and indirectly inhibit the self-replication of CSCs, as CSCs rely on signals from surrounding endothelial cells [114]. This approach can also prevent compensatory mechanisms where hypoxia, induced by other anti-angiogenic therapies, might activate Notch and preserve CSCs [114].

Targeting hypoxia and metabolic vulnerabilities within the TME

Hypoxia is a major obstacle to immunotherapy success, causing immunosuppression and altering immune cell function. Targeting hypoxia and the unique metabolic vulnerabilities of CSCs within the TME is a promising strategy.

• Targeting HIF1α: Hypoxia-inducible factor 1 alpha (HIF1α) signaling is selectively activated in CSCs of hematological malignancies [115]. HIF inhibitors like echinomycin can abrogate the colony-forming unit (cfu) activity of mouse lymphoma and human AML CSCs, leading to preferential elimination of CSCs and efficient eradication in xenogeneic models [115].

• Targeting Hypoxia-induced CBS expression: Studies suggest that targeting hypoxia-induced CBS expression can inhibit breast cancer stem cells through the induction of ferroptosis [116].

• Exploiting Metabolic Vulnerabilities: Understanding the differential metabolic dependencies in tumor versus immune or stromal cells can provide unique therapeutic windows [68]. For example, CSCs often prefer mitochondrial oxidative phosphorylation (OXPHOS) over glycolysis, which confers resistance to glycolysis inhibitors [117]. Targeting these specific metabolic pathways or nutrient competition can impair CSC survival and function [68]. This remains a predominantly preclinical field, as direct HIF inhibitors have struggled in the clinic due to poor drug-like properties, significant toxicities related to HIF’s role in normal tissue, and functional redundancy between HIF-1α and HIF-2α.

Modulating soluble factors and ECM components

The soluble factors and ECM components in the TME create a supportive environment for CSCs, and modulating them can disrupt CSC functions.

• Modulating Secreted Components: Switching the action of pro-cancer secreted components of the TME to anti-cancer activities is a new avenue. This involves targeting various growth factors (e.g., VEGF, HGF), cytokines (e.g., IL-6, IL-8, TGF-β), chemokines, and exosomes that support cancer survival, progression, and drug resistance.12 For instance, neutralizing monoclonal antibodies to IL-8 (e.g., HuMax-IL8) or small-molecule inhibitors blocking CXCR1/CXCR2 (e.g., SX-682) can prevent MDSC recruitment and enhance immunotherapy [51].

• Targeting ECM: The ECM provides structure and support, and its deregulation promotes tumor growth and resistance [118]. Targeting the ECM can improve drug penetration and overcome resistance by reducing fibrosis (e.g., Halofuginone), decreasing collagen crosslinking (e.g., LOX inhibitors), or improving drug penetration (e.g., nab-paclitaxel) [118]. Utilizing mesenchymal stromal cells (MSCs) for drug delivery directly to the tumor site is also a novel strategy [118]. The CXCL12-CXCR4 axis, which ensures close contact between CSCs and tumor stroma, promoting growth, metastasis, and chemoresistance, can be targeted by CXCR4 inhibitors (e.g., AMD3100, CTCE-9908) or CXCL12 inhibitors (e.g., NOX-A12) [119, 120]. These strategies are advancing in clinical trials (Phase I/II), but the primary challenges are the profound redundancy of the TME (e.g., tumors secreting multiple cytokines) and potential systemic toxicities from disrupting fundamental biological processes like ECM homeostasis.

Synergistic and combined therapeutic approaches

The complex and dynamic nature of CSCs and the TME, characterized by heterogeneity, plasticity, and compensatory escape mechanisms, underscores that single-target therapies are often insufficient to achieve durable responses [121]. Therefore, synergistic and combined therapeutic approaches are crucial for overcoming resistance and achieving complete tumor eradication.

• Rationale for Multi-pronged Strategies: Tumor heterogeneity, driven by genetic alterations and epigenetic events, is a major challenge in cancer therapeutics [122]. Tumor cell plasticity, where non-CSCs can de-differentiate into CSCs, further complicates therapies that specifically target CSCs [123]. This necessitates combination therapies that kill both CSCs and non-CSCs, and/or prevent non-CSC-to-CSC transition [123]. The extensive crosstalk among different signaling pathways can lead to compensatory escape, making multitarget inhibitors or combination approaches essential [45].

• Examples of combined therapies:

  • CSC-targeting + TME modulation: Targeting the synergistic interplay between CSCs and their favorable microenvironments is a promising strategy [82]. This involves simultaneously targeting CSC-intrinsic properties (e.g., surface markers, signaling pathways) and disturbing microenvironmental signals (e.g., hypoxia, acidic pH, ECM components, stromal cells) that sustain CSC growth and drug resistance [80]. For example, inhibiting HIF-1α in hypoxic HNSCC CSCs makes them sensitive to radiation [82]. Targeting TGF-β and stromal crosstalk also shows promise for depleting CSCs [82].
  • Immunotherapy + Plasticity Inhibitors: Combining immunotherapy with agents that reduce tumor cell plasticity is a key strategy to overcome resistance [51].

    • IL-8 axis targeting: Neutralizing monoclonal antibodies to IL-8 (e.g., HuMax-IL8) or small-molecule inhibitors blocking CXCR1/CXCR2 (e.g., SX-682) can reduce MDSC recruitment [124] and enhance NK and T cell-mediated killing, alleviating resistance to targeted therapies and sensitizing cells to chemotherapy [51].
    • TGF-β blockade: Small-molecule inhibitors like Galunisertib, or bifunctional fusion proteins like Bintrafusp alfa (anti-PD-L1/TGF-βRII), can revert TGF-β1-induced tumor cell plasticity, enhance anti-tumor responses, and synergize with immune checkpoint inhibitors (ICIs) [51].
    • AXL pathway modulators: Combining ICB with AXL pathway modulators is being explored [51].
    • Cancer vaccines against EMT transcription factors: Novel approaches like Brachyury-based vaccines can activate T cells to eliminate tumor cells expressing these factors, potentially in combination with ICB and EMT-modulating drugs [51, 125].
  • Flavonoids in Combination Treatments: Flavonoids can modulate cell plasticity and re-sensitize resistant breast carcinoma by interfering with the NF-κB pathway and its complex signaling consequences [126]. This includes modifying the TME (containing inflammatory processes, suppressing EMT), regulating growth factor receptors and specific signaling pathways (PI3K/AKT, MAPK, JAK/STAT), modulating ABC transporters, and regulating apoptosis, autophagy, and cell cycle [126–128].
  • Combination with Standard of Care (SoC): Combined therapies with chemotherapy, radiotherapy, and targeted agents are being extensively investigated. For instance, combining targeted therapy with immunotherapy (ICB) is showing feasibility in difficult-to-treat populations, such as in EGFR-TKI-resistant NSCLC [129]. A key example is the Phase III HARMONi-A trial, which demonstrated that combining the PD-1/VEGF bispecific antibody ivonescimab with chemotherapy led to statistically significant improvements in both progression-free and overall survival for this resistant population. Similarly, ICB combined with anti-angiogenic targeted therapy has become a new standard of care in HCC, based on the Phase III IMbrave150 trial; this study showed that atezolizumab (ICB) plus bevacizumab (anti-VEGF) provided superior overall and progression-free survival compared to sorafenib. Chemotherapy plus ICB can also extend synergistic therapeutic effects, as the chemotherapy can induce immunogenic cell death, broadening the immune response. CAR-based immunotherapies are also being combined with chemotherapy, radiotherapy, ICIs, metabolic regulators, anti-VEGF therapy, and signaling pathway inhibitors to enhance efficacy and overcome resistance [130].

Challenges and future directions in clinical translation

Despite significant advancements in understanding CSC plasticity and TME crosstalk, translating these insights into effective clinical practice faces numerous challenges. However, the rapidly evolving landscape of advanced technologies offers promising future directions for precision oncology.

Challenges in translating CSC plasticity and TME crosstalk targeting to the clinic

The inherent complexity and dynamic nature of cancer present significant hurdles for the clinical translation of CSC plasticity and TME crosstalk targeting strategies.

• Tumor Heterogeneity and Dynamic Phenotype Switching: Cancer exhibits profound cellular diversity, with variations within a single tumor (intratumoral) and between patients (intertumoral) [122]. This heterogeneity, generated by genetic alterations and epigenetic events, means that different cells in a tumor may have different mutations, making it challenging to administer a drug combination that effectively targets every cancer cell [123]. Furthermore, cancer cells display remarkable phenotypic plasticity, allowing them to polarize towards drug-resistant phenotypes [130]. non-CSCs can spontaneously de-differentiate into CSCs, replenishing the CSC pool even if existing CSCs are eliminated, leading to tumor relapse and acquired resistance [123]. This dynamic switching makes CSCs a “moving target” [117].

• Lack of Universal CSC Markers and Reliable Biomarkers for Plasticity: A key challenge is the lack of universal markers for CSCs, as CSCs and normal stem cells can express the same biomarkers, complicating the isolation of pure tumorigenic populations [42]. The elevated levels of plasticity in CSCs further alter their functional phenotype and appearance in response to therapies, making it difficult to find reliable biomarkers for therapeutic application [131]. This non-specificity of CSC signaling limits the identification of appropriate therapeutic targets [132].

• Drug Delivery and Penetration Barriers within the TME: The TME itself poses significant physical and physiological barriers to drug delivery and penetration. Hypoxia, acidic pH, and a dense, fibrotic extracellular matrix (ECM) create a physical barrier that reduces drug accessibility and diffusion [133]. Defective blood vessels within the tumor also impede the effective delivery of therapeutic drugs [133]. These challenges are particularly pronounced for larger molecules like antibodies or cell-based therapies [134]. Overcoming these barriers is a key focus of nanomedicine. For example, Zhang et al. have developed multifunctional magnetic Prussian blue nanoplatforms that integrate imaging and chemo-photothermal therapy, designed specifically to enhance drug penetration and efficacy within the complex TME [135].

• Immune Suppression and Evasion Mechanisms: CSCs possess intrinsic mechanisms to evade immune surveillance, including low expression of major histocompatibility complex (MHC) molecules, upregulation of immune checkpoint proteins (e.g., PD-L1, CD47), and secretion of immunosuppressive cytokines [11]. The TME actively contributes to this immune suppression by recruiting immunosuppressive cells (e.g., Tregs, MDSCs, TAMs) and creating an environment that dampens anti-tumor immune responses and excludes T cells [75]. Immunotherapy itself can paradoxically drive stem cell-like properties in tumors, leading to an “immune resistance continuum” [117]. Given the bidirectional plasticity of the immune milieu, severe immune dysregulations can emerge; a recent case documents autoimmune encephalitis followed by HLH (Huang et al. [136]), underscoring vigilance during immune-modulating regimens.

• Off-target Toxicities and Compensatory Pathways: Many CSC-targeting agents lack specificity and can affect healthy tissue stem cells, leading to undesirable side effects [45]. For instance, approved Smo inhibitors have ubiquitous toxic side effects that limit their application [45]. The complex crosstalk among different signaling pathways can lead to compensatory escape mechanisms, where inhibition of one pathway might upregulate another, rendering single-target inhibitors ineffective [45].

• Limitations of Preclinical Models and Clinical Trial Design: Most CSC studies are performed in immune-deficient mice, which do not fully recapitulate the biological complexity of human tumors, including the crucial immune microenvironment [42]. The small number of CSCs that can be isolated and kept functional limits high-throughput screening in tumor models [137]. Furthermore, current clinical trial designs often use traditional endpoints like tumor shrinkage, which are insufficient to evaluate the efficacy of CSC-targeting agents that may primarily affect stemness or dormancy rather than immediate tumor size [122]. The difficulty in predicting which patients will respond to specific therapies leads to many patients receiving toxic and expensive treatments without benefit [137].

Future perspectives and the era of precision oncology

Overcoming the formidable challenges in targeting CSC plasticity and TME crosstalk requires a concerted effort leveraging advanced technologies and a deeper understanding of tumor biology. The future of cancer therapy lies in the era of precision oncology, driven by innovative approaches.

• Leveraging Advanced Technologies: Cutting-edge technologies are crucial for elucidating the complex interplay within tumors.

  • Single-cell Multiomics and Spatial Transcriptomics: These technologies allow for the elucidation of the genome, epigenome, and transcriptome at a single-cell level, providing unprecedented resolution into intratumoral heterogeneity and dynamic cellular states [76]. Spatial transcriptomics further reveals the spatial distribution of cell types and their interactions within the TME, offering crucial insights into the immune landscape and identifying promising targets for immunotherapy [76].
  • Artificial Intelligence (AI) and Machine Learning: AI can assist in analyzing TME characteristics and patient demographics to predict treatment response [130]. Machine learning can identify design principles for CARs with high cytotoxicity and persistence, and AI-based methods can improve patient follow-up and response assessment by distinguishing between pseudo-progression and true-progression [130].
  • Organ-on-chip platforms and 3D Bioprinting: These advanced in vitro models can mimic the cellular composition of the TME and provide biomimetic body fluid, maintaining tumor cell phenotypes and tumor grade for personalized therapeutics [137]. They offer well-controlled studies of cell-cell interactions, CSC behavior, and CSC-specific biomarkers, facilitating drug screening and understanding of CSC response to cancer drugs [137].

• Development of Novel Biomarkers for Plasticity, Response, and Resistance: Identifying novel and precise biomarkers for CSC subpopulations and their dynamic changes is crucial for developing more targeted therapies and for real-time tracking of treatment effects [131]. These biomarkers will be essential for guiding personalized treatments and assessing therapeutic efficacy beyond traditional tumor shrinkage metrics [122].

• Rational Design of Adaptive and Personalized Therapies: The understanding that cancer cells can adapt to therapeutic pressure by shifting phenotypes necessitates the rational design of adaptive and personalized therapies [27]. This involves tailoring treatments based on individual patient characteristics and tumor biology, potentially using intermittent dosing schedules or concurrent suppression of pivotal pathways to prevent drug-tolerant cells from reaching an irreversibly resistant state [138]. The integration of multiomics data will enable the specification of sensitive cancer types based on individual patient characteristics [126].

• Exploiting Metabolic Vulnerabilities for Therapeutic Gain: A deeper understanding of the metabolic reprogramming of immune cells within the TME and the distinct metabolic characteristics of CSCs can reveal novel opportunities for immunotherapeutic approaches [117]. By targeting the metabolic states of immune cells, it is possible to augment their anti-tumor activities, and making CSCs more vulnerable to immune elimination by targeting their metabolic plasticity holds significant promise [117].

• Stem Cell-Based Strategies: Stem cell therapy offers a hopeful option by improving therapeutic efficacy and reducing off-target events due to enhanced targeting of tumors [83]. Concurrently, clinical maturation of lentiviral gene therapy in pediatric β0/β0 transfusion-dependent β-thalassemia (Li et al. [139]) highlights the safety/efficacy frameworks transferrable to next-gen CSC-targeting cellular products.

• This includes HSC transplantation [140], MSC transplantation, stem cells as therapeutic carriers (genetically modified stem cells, nanoparticles-carrying stem cells [141], oncolytic virus carriers [142], exosome carriers), and stem cell sources for immune cell production (CAR T cells, NK cells) [83, 143]. Stem cell-based anti-cancer vaccines targeting CSCs are also being explored [83, 144]. This field is rapidly advancing, with novel materials like metal–organic frameworks (MOFs) being engineered as sophisticated carriers for targeted anti-cancer drug delivery, offering new avenues to tackle both delivery and penetration challenges [145].

• TME Normalization: Strategies to normalize the TME, including vascular normalization and reprogramming of CAFs, can enhance anti-tumor immunity and immunotherapy efficacy by alleviating hypoxia and reducing physical barriers to immune cell migration [146]. This creates a more permissive environment for immune cells to infiltrate and function, thereby enhancing anti-tumor immunity [146].

Conclusion

Cancer’s remarkable adaptability and therapeutic resistance are not driven by a static tumor hierarchy but by the dynamic interplay between cancer stem cell (CSC) plasticity and the tumor microenvironment (TME). This review establishes that the TME is an active accomplice, engaging in bidirectional crosstalk that sustains stemness, fuels progression, and facilitates immune evasion. This symbiotic relationship, orchestrated by complex signaling, epigenetic, and metabolic mechanisms, is a central determinant of therapeutic failure.

While therapeutic strategies targeting specific CSC-intrinsic pathways (e.g., Wnt, Notch) or TME components (e.g., CAFs, TAMs) represent significant progress, their efficacy is often transient. As this review highlights, the profound challenges of tumor heterogeneity, rapid phenotypic switching, and a deeply immunosuppressive microenvironment mean that single-target interventions are insufficient. The key therapeutic implication is the necessity of multi-pronged, synergistic approaches that can simultaneously disrupt the CSC’s internal survival programs and dismantle its external support niche.

The future of cancer therapy lies in a holistic approach rooted in precision oncology. This vision is now attainable through advanced technologies like single-cell multiomics, spatial transcriptomics, and artificial intelligence, which can deconstruct tumor evolution in real-time. The rational design of adaptive, personalized therapies—combining direct CSC targeting, TME normalization, and immunotherapy—holds the key to overcoming resistance and achieving durable patient benefits.

Author contributions

Mutaz Jamal Al-khreisat, Waleed K. Abdulsahib, and Ihsan Khudhair Jasim contributed to the conception and design of the manuscript. H. Malathi, Priya Priyadarshini Nayak, and D. Alex Anand contributed to the literature review and analysis. Gunjan Mukherjee, Aashna Sinha, and Oybek Ruziyev contributed to drafting, critical revision, and editing of the manuscript. All authors read and approved the final version of the manuscript and agree to be accountable for all aspects of the work.

Funding

No funding was received for this work.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.