Cell reprogramming in cancer: Interplay of genetic, epigenetic mechanisms, and the tumor microenvironment in carcinogenesis and metastasis

Abstract

Cell plasticity, also known as lineage plasticity, refers to the ability of a cell to reprogram and change its phenotypic identity in response to various cues. This phenomenon is context-dependent, playing a crucial role in embryonic development, tissue regeneration, and wound healing. However, when dysregulated, cell plasticity contributes to cancer initiation, progression, metastasis, and therapeutic resistance. Throughout different stages of tumor development, cancer cells exploit various forms of plasticity to evade normal regulatory mechanisms that govern cell division and homeostasis. Recent evidence highlights the complex interplay between genetic and epigenetic factors, the tumor microenvironment, and epithelial-to-mesenchymal transition in driving cancer cell plasticity. This dynamic reprogramming suggests that “deregulated cell plasticity” could be considered an additional hallmark of cancer. Advancements in next-generation sequencing and single-cell RNA analysis, combined with artificial intelligence technologies such as deep learning, along with Google’s AlphaFold may help predict the trajectories of cancer cells. By predicting protein three-dimensional structures and identifying both active and potential allosteric binding sites, AlphaFold 2 can accelerate the development of new cancer drugs and therapies. For example, allosteric drugs, bind to the allosteric rather than the active sites, can induce conformational changes in proteins, affecting their activities. This can then alter the conformation of an active site that a drug-resistant mutation has created, permitting a blocked orthosteric drug to bind and this enables the design of more effective drugs that can synergize with traditional orthosteric drugs to bind and regain its efficacy. These innovations could provide deeper insights into the intricate mechanisms of cancer progression and resistance, ultimately paving the way for more precise, durable, and personalized oncologic treatments.

Key Words: Cell reprogramming; Tumorigenesis; Chemotherapy resistance; Artificial intelligence; Deep learning; AlphaFold

Core Tip: The idea of deregulated cell plasticity being considered a hallmark of cancer emphasizes the adaptive nature of tumors, their ability to evade homeostatic regulation, and their capacity for resilience in the face of therapeutic intervention. Understanding the cell reprogramming mechanisms that drive this plasticity could ultimately lead to the development of more effective therapies that target these adaptive processes and prevent the tumor from exploiting its ability to continuously evolve and resist treatment. Google’s AlphaFold represents a game-changer in chemotherapy drug discovery by providing a fast and accurate method to predict protein structures, enabling researchers to identify new drug targets, design better molecules, and understand the complex biology of cancer at a level of detail that was previously unattainable. By integrating next-generation sequencing and single-cell RNA sequencing, artificial intelligence-driven deep learning models, and AlphaFold’s predictions with experimental data and existing drug discovery techniques, we could see faster development of more effective, specific, chemotherapy agents in the future. By merging these technologies, precise, durable, and truly personalized cancer therapy could become a clinical reality, revolutionizing the way we approach cancer treatment.

INTRODUCTION

Cancer has often been described as a wound that does not heal, a concept initially proposed by Virchow and later articulated by Dvorak[1]. Carcinogenesis begins with an oncogenic insult after a driver mutation grants growth advantage to a cell within tissues. Subsequent proliferation of this initial clone, accompanied by additional somatic mutations, epigenetic modifications, and interactions with tumor microenvironment (TME), leads to heterogeneity and complexity observed in advanced tumors and ultimately responsible for cancer-related deaths. However, before reaching an advanced stage, tumor cells must undergo several intermediate steps, including invasion, intravasation, survival in circulation, extravasation, micro-metastasis, and colonization. These processes require cancer cells to reprogram and alter their phenotypic identity i.e., demonstrating plasticity to adapt, metastasize, and thrive in newer environments.

The idea of deregulated cell plasticity being considered a hallmark of cancer emphasizes the adaptive nature of tumors, their ability to evade homeostatic regulation, and their capacity for resilience in the face of therapeutic intervention. Understanding the cell reprogramming mechanisms that drive this plasticity could ultimately lead to the development of more effective therapies that target these adaptive processes and prevent the tumor from exploiting its ability to continuously evolve and resist treatment. The foundational understanding of cancer biology was significantly advanced by Hanahan and Weinberg’s landmark framework on the Hallmarks of Cancer in 2000[2], later updated in 2011[3]. As cancer research progresses, these hallmarks continue to evolve, incorporating new insights into tumor biology. Here, we discuss the experimental evidence and explore role of cell plasticity in cancer initiation, progression, metastasis, and therapeutic resistance.

CELL PLASTICITY: DEFINITIONS, ROLE IN EMBRYONIC DEVELOPMENT AND TISSUE HOMEOSTASIS

Cell plasticity refers to the ability of a cell to reprogram and change its phenotypic identity. This phenomenon plays a crucial role in embryonic development, tissue regeneration, wound healing, and cancer progression[4-6]. Cell reprogramming can occur through distinct mechanisms (Figure 1). The first, de-differentiation, involves a specialized cell reverting to a less specialized state within the same lineage, potentially regaining stem-like properties. For example, in colorectal neoplasia, cancer stem cell (CSC)-like properties may arise from progenitor cells via de-differentiation, where committed cells lose their specialized identity and reacquire stem-like features. The second, trans-differentiation, describes a cell transitioning from one differentiated type to another, often across lineages. Example include, the phenomenon of epithelial-to-mesenchymal transition (EMT). Tumor cells can transition from epithelial cells to cancer-associated myofibroblasts and later revert back to epithelial states through mesenchymal-to-epithelial transition (MET) upon metastasis and colonization. The third, trans-determination, refers to conversion of a progenitor stem cell population into another, forming the basis of metaplasia as noted in Barrett’s esophagus in chronic acid exposed distal esophagus which may lead to esophageal cancer[4,5,7].

Figure 1 This schematic highlights the dynamic plasticity of cancer cells and their ability to reprogram, contributing to tumor progression and therapy resistance. Cancer initiation begins with a cell of origin, which is the first cell in a tissue to acquire a cancer-initiating driver mutation. This cell can be a normal native stem cell, a common/committed progenitor, or a fully differentiated cell. Driver mutations (indicated by red starbursts) can occur at any stage of cell differentiation, leading to malignant transformation. The result is a heterogenous tumor consisting of cell of origin, cancer stem cells, blood vessels, immune cells. De-differentiation refers to the reprogramming of a specialized cell into a less specialized state, reverting to a progenitor or stem-like phenotype within the same lineage. Trans-differentiation occurs when a differentiated cell reprograms into another differentiated cell type, potentially from a different lineage, as observed in treatment-induced neuroendocrine prostate cancer.

During normal embryonic development, cell plasticity follows a unidirectional and irreversible progression from a totipotent zygote to a fully differentiated state, metaphorically described by Waddington[8] as a ball rolling down a hill, where specific signaling events progressively restrict cell fate[6,9]. For instance, EMT is essential for mesoderm and endoderm formation during gastrulation, as well as for neural crest cell migration and fate specification[6,10]. EMT allows epithelial cells to lose polarity and adhesion, acquiring mesenchymal properties and migratory potential, mediated by transcription factors such as Slug, Snail, Zeb, and Twist. Notably, these factors are reactivated in cancer, driving metastasis[5,11].

Traditionally, terminally differentiated cells were considered post-mitotic, incapable of re-entering the cell cycle. This stability is maintained through epigenetic mechanisms, including DNA methylation, polycomb repressors, and chromatin remodeling, which silence stem cell factors and preserve lineage fidelity[12]. However, research has shown that under certain conditions, differentiated cells can reprogram and re-enter the cell cycle. Classic experiments in amphibian developmental biology demonstrated that nuclei from differentiated cells, when introduced into enucleated Xenopus eggs, could reset their transcriptional activity to an early embryonic state, highlighting the retained plasticity of differentiated nuclei[13].

In the context of cancer, these findings suggest that cell plasticity - manifesting as de-differentiation, trans-differentiation, and EMT - may be influenced by signals from TME. The tumor milieu, through autocrine, juxtacrine, and paracrine signaling, can drive metabolic reprogramming, metastasis, and therapy resistance[11]. In adulthood, cell plasticity contributes to tissue homeostasis, responding to physiological stressors such as injury, inflammation, and senescence, facilitating wound healing and regeneration[8,14,15]. For example, in the intestinal crypts, when leucine-rich-repeat-containing G-protein-coupled receptor 5 (LGR5)-positive stem cells are ablated, lost cells are replenished through the de-differentiation of transit-amplifying progenitors or committed secretory cells[16,17]. Similar injury-induced plasticity has been observed in hepatocyte repair[18,19], the pancreas[20,21], and the prostate[22,23].

While cell plasticity is essential for regeneration, it can also contribute to pathological conditions such as metaplasia. In chronic tissue injury, plasticity-driven lineage conversion can lead to conditions such as Barrett’s esophagus, where esophageal squamous epithelium transforms into an intestinal-like columnar phenotype[24-26]. This underscores the contextual nature of cell plasticity, where tightly regulated signaling promotes regeneration in some cases but drives disease progression in others.

CELL PLASTICITY: CARCINOGENESIS AND METASTASIS

Cell plasticity, when deregulated plays a crucial role in tumor initiation, progression, metastasis, and resistance to therapy[4,6]. It is important to distinguish between the cell of origin and the CSC. The cell of origin refers to the first cell within a tissue that acquires the driver mutation, also known as the tumor-initiating cell[27]. This cell may be a native stem cell of the tissue (hierarchical model) or a committed progenitor or differentiated cell (stochastic model)[14]. Examples include: LGR5+ intestinal crypt stem cells in colorectal cancer[28,29], long-term progenitor cells in the interfollicular epidermis and upper infundibulum for basal cell carcinoma[30]. Both stem and progenitor cells in pediatric cerebellar tumors[31] and acinar cells in pancreatic cancer, undergoing trans-differentiation (acinar-to-ductal metaplasia), particularly in the setting of chronic pancreatitis[32,33]. By contrast, CSCs represent a subset of tumor cells with the ability to self-renew and differentiate into multiple lineages, sustaining and propagating tumor growth. Importantly, CSCs may possess different molecular and phenotypic characteristics than the original cell of origin[14,27,28,34].

Cancer cells do not maintain a fixed hierarchical structure but rather evolve dynamically, exhibiting early genetic heterogeneity that drives subsequent plasticity. One of the fundamental mechanisms underlying tumor evolution is reactivation of cellular reprogramming machinery, enabling unrestricted proliferation, varying potency, and differentiation into either the original lineage or multiple distinct epithelial lineages - essentially adopting a CSC/progenitor-like state[3,4,14,28].

This phenotypic plasticity grants tumor cells a growth and survival advantage, contributing to the heterogeneity observed in advanced tumors. Notably, CSCs are defined by their ability to initiate and sustain tumor formation upon transplantation into recipient hosts[3]. The debate continues regarding the precise origin of CSCs, whether they arise directly from normal stem cells or from differentiated/progenitor cells that undergo de-differentiation. In colorectal neoplasms, evidence suggests an origin in long-lived intestinal stem cells, where stochastic oncogenic mutations in LGR5+ crypt cells drive CSC formation. For instance, an adenomatous polyposis coli (APC) mutation disrupts Wnt-β catenin signaling, leading to unchecked proliferation and the formation of aberrant crypt foci, the precursors to adenomas and eventually carcinomas. However, the same APC mutation in transit-amplifying cells only results in micro-adenomas, insufficient for full malignancy[14,27-29]. Alternatively, CSC-like properties may arise from progenitor cells via de-differentiation, where committed cells lose their specialized identity and reacquire stem-like morphology[4,6,28,35,36].

Beyond de-differentiation, trans-differentiation may drive tumor plasticity, influencing progression, particularly through EMT. Tumor cells can transition from epithelial cells to cancer-associated myofibroblasts and later revert back to epithelial states through MET upon metastasis and colonization. This process is influenced by signals from the TME, including tumor-associated macrophages, fibroblasts, endothelial cells, and pericytes, which contribute to epigenetic reprogramming and altered gene expression[4,5,15,28]. In addition to oncogenic drivers, loss of tumor suppressor mechanisms further facilitates this plasticity. By understanding the mechanisms underlying cell plasticity in tumors, new therapeutic strategies may emerge, targeting cancer progression and resistance at its root.

ROLE OF ONCOGENES, TUMOR SUPPRESSOR’S GENES, IN CELL PLASTICITY

The interplay between oncogenes, transcription factors, homeobox genes, tumor suppressors, and the TME creates a dynamic state of cell plasticity that fuels cancer initiation, progression, and resistance. Understanding these mechanisms may offer new therapeutic targets to disrupt tumor evolution and improve treatment outcomes (Table 1).

Table 1 Summary of commonly described genes, epigenetic regulators, tumor microenvironment factors, and epithelial-mesenchymal transition factors involved in cell reprogramming and cancer cell plasticity.

	Summary of commonly described genes
Key genes in cell reprogramming and cancer plasticity	Oncogenes: KRAS, MYC, EGFR, MET, PIK3CA, BRAF
	Tumor suppressor genes: TP53, RB1, PTEN, CDKN2A, APC
	Stemness/reprogramming genes: OCT4, SOX2, NANOG, KLF4, LIN28, c-MYC
	Neuroendocrine differentiation genes: ASCL1, NEUROD1, DLL3, CHGA, SYP
	EMT-associated transcription factors: SNAIL (SNAI1), SLUG (SNAI2), TWIST1, ZEB1, ZEB2
Epigenetic regulators	Histone modifiers: EZH2, KDM6B, SETDB1, NSD2 (regulate chromatin accessibility)
	DNA methylation enzymes: DNMT1, DNMT3A, DNMT3B, TET1-3 (control gene silencing or activation)
	Chromatin remodelers: SWI/SNF complex (ARID1A, BRG1/SMARCA4) (alter chromatin structure)
	Non-coding RNAs: MiR-200 family, lncRNAs (HOTAIR, MALAT1) (modulate gene expression)
Tumor microenvironment factors	Cytokines & chemokines: IL-6, IL-8, TGF-β, TNF-α, CXCL12, CCL2 (promote inflammation and plasticity)
	Immune cells: Tumor-associated macrophages, myeloid-derived suppressor cells, T-regulatory cells (contribute to immune evasion)
	Hypoxia-induced factors: HIF-1α, VEGF, ANGPT2 (promote angiogenesis and plasticity)
	Fibroblasts & extracellular matrix: Cancer-associated fibroblasts, matrix stiffness (collagen I, fibronectin, hyaluronan)
EMT factors in cancer cell plasticity	Transcription factors: SNAIL, SLUG, TWIST1, ZEB1, ZEB2 (suppress epithelial markers and enhance mesenchymal transition)
	Epithelial markers (downregulated in EMT): CDH1, cytokeratin (KRTs), CLDNs, OCLN
	Mesenchymal markers (upregulated in EMT): VIM, CDH2, FN1
	Signaling pathways driving EMT: TGF-β, WNT/β-catenin, Notch, Hedgehog, PI3K/AKT, NF-κB

EGFR: Epidermal growth factor receptor; BRAF: B-Raf proto-oncogene; PTEN: Phosphatase and tensin homolog; CDKN2A: Cyclin-dependent kinase inhibitor 2A; APC: Adenomatous polyposis coli; OCT4: Octamer-binding transcription factor 4; SOX2: Sex-determining region Y-box 2; KLF4: Kruppel-like factor 4; DLL3: Delta-like ligand 3; CHGA: Chromogranin A; SYP: Synaptophysin; ZEB1: Zinc finger E-box binding homeobox 1; ZEB2: Zinc finger E-box binding homeobox 2; EZH2: Enhancer of zeste homolog 2; SETDB1: SET domain bifurcated 1; NSD2: Nuclear SET domain 2; DNMT: DNA methyltransferase; TET: Ten-eleven translocation; SWI/SNF: SWItch/sucrose nonfermentable; ARID1A: AT-rich interaction domain 1A; BRG1: Brahma-related gene 1; lncRNA: Long noncoding RNA; HOTAIR: HOX transcript antisense RNA; MALAT1: Metastasis associated lung adenocarcinoma transcript 1; IL: Interleukin; TGF: Transforming growth factor; TNF: Tumor necrosis factor; CXCL12: C-X-C motif ligand 12; CCL2: C-C motif ligand 2; HIF-α: Hypoxia inducible factor alpha; VEGF: Vascular endothelial growth factor; ANGPT2: Angiopoietin 2; CDH1: E-cadherin; CLDNs: Claudins; OCLN: Occludins; VIM: Vimentin; CDH2: N-cadherin; FN1: Fibronectin 1; EMT: Epithelial-mesenchymal transition; PI3K: Phosphatidylinositol-3-kinase; AKT: Protein kinase B; NF-κB: Nuclear factor-kappa B.

Cell plasticity is influenced by both cell-autonomous (intrinsic) and non-cell-autonomous (extrinsic) factors[6,27,28]. Intrinsic factors include somatic mutations with DNA damage, epigenetic deregulation with DNA methylation, and cellular senescence. Extrinsic factors could be secondary to infections, inflammation and injury, interactions of the immune cells with the TME and responses to drug treatment. A variety of signaling pathways due to growth factors contribute to plasticity, including transforming growth factor-β, bone morphogenetic protein, WNT, fibroblast growth factor, hepatocyte growth factor, platelet-derived growth factor, Notch, sonic hedgehog, vascular endothelial growth factor, and hypoxia inducible factor alpha (HIF-α). Together, these elements promote tumor initiation, progression, and survival[3,6,15,28]. Early evidence of intrinsic cellular reprogramming came from studies showing that a single gene (MyoD) could induce fibroblast-to-muscle cell trans-differentiation[37]. Later, Takahashi and Yamanaka[38] in 2006 identified a set of four transcription factors - octamer-binding transcription factor 4 (Oct4), sex-determining region Y-box 2 (Sox2), Kruppel-like factor 4, and c-Myc - capable of converting fibroblasts into induced pluripotent stem cells, demonstrating de-differentiation.

While these transcription factors regulate normal development and tissue regeneration, their misexpression in cancerous contexts can drive tumorigenesis. For example mechanism includes: Oct4 inhibits differentiation, maintains proliferation, and induces intestinal dysplasia[39], Kruppel-like factor 4 cooperates with KRAS mutations to induce acinar-to-ductal trans-differentiation, promoting pancreatic cancer[40], Sox2 enhances reprogramming in prostate, lung, and pancreatic cancers, as well as EMT[6,41,42], and Myc a potent oncogene frequently overexpressed in cancers, implicated in pediatric neuroblastomas [in cooperation with anaplastic lymphoma kinase (ALK) mutations] and EMT induction in pancreatic cancer alongside KRAS mutations[43,44].

Reprogramming capabilities may arise through oncogenic mutations (gain of function), transcription factor amplification, translocation, or tumor suppressor mutation (loss-of function), often reinforced by epigenetic modifications[6,45]. Over-activation of stem cell pathways (MYC, WNT, epidermal growth factor, transforming growth factor-β, Notch, Hedgehog), combined with mutations in epidermal growth factor receptor (EGFR), KRAS or inactivation of tumor suppressors [RB1, TP53, APC, BRCA, phosphatase and tensin homolog (PTEN)] contributes to carcinogenesis[4-6,28]. Other relevant transcriptional genes which specify cell fate are homeobox genes. Homeobox genes are evolutionarily conserved transcriptional regulators that control cell fate and tissue identity during development. Their expression is often retained in cancers and used diagnostically to determine tumor origin in cases where the primary site is unknown[6,21,25,46,47]. Examples include: Caudal-related homeobox transcription factor 2 in colon/intestines, NK2 homeobox 1 (NKX2.1) in lung, pancreatic and duodenal homeobox 1 in pancreas, paired box 4 in pancreas, SOX in pancreas/prostate, NKX3.1 in prostate, microphthalmia-associated transcription factor in melanocytes.

Dysregulated expression of homeobox genes in the context of chronic inflammation, injury, and metaplasia can drive reprogramming and plasticity in cancer[6,21,25,46,47]. In rat lung cancer models, thyroid transcription factor 1 (NKX2.1) loss combined with KRAS activation led to mucinous adenocarcinomas resembling gastric/intestinal tumors. Conversely, NKX2.1 Loss with SOX2 overexpression resulted in a squamous phenotype with esophageal differentiation[48]. A glioblastoma study highlighted how genetic alterations in cyclin-dependent kinase 4, EGFR, platelet-derived growth receptor alpha, neurofibromatosis type 1 interact with the TME to shape tumor phenotype and cellular states through plasticity[49].

Similar to oncogenes and transcription factors as described, tumor suppressors also play a role in tumor initiation and progression. Tumor suppressors such as TP53, RB1, and PTEN are critical in preventing uncontrolled cell proliferation. Their loss, whether by mutation, deletion, or epigenetic silencing, removes key restraints on oncogenic activity, allowing for reprogramming and plasticity[6,33,45]. For example, RB1 directly represses SOX2, Oct4, and Nanog, preventing reprogramming. RB1 Loss, therefore, enhances cellular plasticity. High-risk human papillomavirus infections inhibit TP53 (via E6 protein) and RB1 (via E7 protein), enabling unrestricted proliferation, as seen in human papillomavirus-driven cervical and oropharyngeal cancers[2,3]. PTEN loss in prostate cancer facilitates trans-differentiation from basal to luminal cells, contributing to tumor progression[6,33,45].

EPIGENETIC MODULATION IN CANCER PLASTICITY AND TUMORIGENESIS

Recent advances in genome sequencing have highlighted the role of epigenetic deregulation in carcinogenesis, metastasis, and CSC induction. Aberrant epigenetic modifications can make chromatin either too restrictive, leading to differentiation blocks, or overly permissive, enabling aberrant cell reprogramming[3,4,12,45]. One hallmark of epigenetic dysregulation in cancer is the reactivation of bivalent chromatin state, characterized by both active (H3K4me3) and repressive (H3K27me3) histone marks at transcription factor promoters. This phenomenon mirrors embryonic development, where the chromatin remains poised for transcription in dividing and differentiating cells but stabilizes upon terminal differentiation. However, in cancer, the dysregulation of bivalent chromatin contributes to tumorigenesis by disrupting normal cell fate decisions and maintaining an aberrant plastic state. In CSCs, persistent bivalency facilitates continuous cell reprogramming and phenotypic switching, promoting unchecked growth and proliferation[3,4,12,45]. Aberrant regulation of histone-modifying enzymes, such as enhancer of zeste homolog 2 (EZH2) (which deposits H3K27me3) and MLL1/2 (which deposit H3K4me3), can lead to the expansion of bivalent chromatin states in cancer. Cancer cells can exploit bivalent chromatin states to survive under therapeutic pressure, leading to drug resistance. For example, in breast and prostate cancers EZH2 overexpression disrupts bivalent chromatin regulation, facilitating tumor progression[3,4,12,45]. Epigenetic modifications contributing to cancer progression include: DNA methylation (hypermethylation or hypomethylation), histone modifications (methylation, acetylation), chromatin remodeling (via the SWItch/sucrose nonfermentable complex) (Table 1).

These alterations can arise through genetic mutations in epigenetic regulators or non-mutagenic stimuli[12,45]. Mutations in epi-genes include, DNA methylation, chromatin remodeling, and histone modification have been implicated in tumorigenesis[50,51]. Certain examples include: DNA methylation enzymes: DNA methyltransferase 1, DNA methyltransferase 3A (methylation); ten-eleven translocation (TET) family (demethylation), chromatin remodelers: AT-rich interaction domain 1A (ARID1A), ARID1B, ARID2, cohesion complex, histone modifiers: H3.3, H3F3A, alpha-thalassemia mental retardation X-linked, DAXX, HIST1H3B, HIST1H1C, histone acetylators/methylators: Histone acetyltransferase, KDM, polycomb repressors: EZH2, insulators: CCCTC-binding factor (CTCF) (Table 1).

Histone modifications are key factors in chromatin packaging responsible for gene regulation during cell fate determination and development. Abnormal alterations in histone modifications potentially affect the stability of the genome and disrupt gene expression patterns leading to cancer. Histone acetylation also plays a role in promoting cancer cell plasticity. This post-translational modification is primarily controlled by histone acetyltransferases, which add acetyl groups to lysine residues on histone tails, and histone deacetylases, which remove them. For example, in glioblastoma, increased histone acetylation at SOX2 and OCT4 promoters enhances their expression, driving tumor cell reprogramming toward a neural stem cell-like state enabling dedifferentiation. Similarly in breast and lung cancers, hyperacetylation at the zinc finger E-box binding homeobox 1 (ZEB1) promoter enhances EMT and metastatic potential. Histone deacetylases inhibitors can reverse EMT by repressing ZEB1 and restoring E-cadherin expression. In EGFR-mutant non-small cell lung cancer (NSCLC) treated with tyrosine kinase inhibitors, histone acetylation can drive trans-differentiation into a neuroendocrine phenotype (small cell lung cancer), leading to resistance. Similarly in castration-resistant prostate cancer (CRPC) undergoing androgen receptor (AR)-targeted therapy, tumor can transition into neuroendocrine prostate cancer, with histone acetylation at ASCL1 and SOX2 loci contributing to lineage switching. Histone deacetylases inhibitors (e.g., vorinostat, panobinostat) can reverse plasticity-driven resistance by reprogramming cancer cells toward a more differentiated, therapy-sensitive state. Combination strategies with epigenetic drugs + targeted therapies (e.g., histone deacetylases inhibitors + AR inhibitors in prostate cancer) are being explored to counteract lineage plasticity[52].

Hypermethylation of CpG-rich promoter regions in tumor suppressor genes and global hypomethylation of oncogenes are frequently observed in cancers[12,28,53,54]. Examples include: DNA hypermethylation in retinoblastoma (RB1) gene promoter[55], Von Hippel-Lindau hypermethylation in renal cell carcinoma[56], hypermethylation patterns in BRCA (breast cancer) and MMR genes (colon cancer)[57-59] and P16 (cyclin-dependent kinase inhibitor 2A) hypermethylation is a common event in cancers, leading to cell cycle dysregulation and evasion of senescence[60].

Similarly, tumor suppressors such as TET assist with oxidation and demethylation of 5-methyl-cytosine (CpG nucleotide) and promote locus specific reversal of DNA methylation thus preventing cellular transformation. Suppressed TET activity, often due to promoter hypermethylation, promotes uncontrolled proliferation and has been observed in gliomas, acute myeloid leukemia, and renal cancers[12,28,61]. TET loss is also linked to hypoxia, HIF-α activation, and isocitrate dehydrogenase 1/2 mutations[50,62]. While tumor suppressor genes undergo hypermethylation and silencing, oncogene activation is frequently associated with hypomethylation (demethylation). This has been observed in BCL2, HIF, and HOX11, contributing to various malignancies[63-65].

Advances in cancer epigenetics have revealed alterations in chromatin conformation and disruptions in genome organization. One critical mechanism involves DNA hypermethylation leading to the disruption of chromatin insulators such as CTCF by the following mechanisms[12]. Topologically associated domains are functional genomic units, bound by CTCF and cohesins at their bases, which constrain transcriptional activity. Hypermethylation of CTCF genes disrupts topologically associated domain integrity, leading to aberrant oncogene activation. Loss of CTCF activity has been implicated in succinate dehydrogenase-deficient tumors and pediatric gliomas[12,63,66].

Enhancers regulate gene expression via long-range chromatin interactions and are decorated with transcription factor motifs, histone modifiers, and chromatin remodeling factors. Alterations in enhancer landscapes have been implicated in cell plasticity and therapy resistance[12,66-68]. Epigenetic dysregulation, thus, plays a crucial role in cancer initiation, plasticity, and resistance by modulating chromatin accessibility, transcriptional states, genome organization and metabolic reprogramming. Targeting epigenetic modifiers and chromatin remodeling pathways represents a promising therapeutic approach for overcoming tumor heterogeneity and treatment resistance.

Recent discoveries in cancer metabolism highlight how disruptions in metabolites and metabolic enzymes influence epigenetic regulation. Dysregulated metabolism enables tumor cells to produce essential building blocks while also influencing epigenetic modifications to drive cancer initiation and progression. Cancer-associated metabolic shifts may also reshape the epigenetic landscape, particularly through alterations in histone and DNA modifications, fostering malignant transformation, adaptation to nutrient scarcity, and metastasis. The accumulation of certain metabolites in cancer can target epigenetic enzymes to globally alter the epigenetic landscape. This metabolically driven reprogramming of epigenetics in cancer is highly dependent on nutrient availability for tumor cells[69].

ROLE OF TME IN CELL PLASTICITY

The TME plays a central role in cancer plasticity, influencing tumor initiation, EMT-driven metastasis, and therapy resistance. Through chronic inflammation, metabolic rearrangements, stromal remodeling, extracellular matrix stiffening, and systemic instigation, the TME fosters an adaptive environment that supports tumor evolution and immune evasion. Understanding these mechanisms offers new therapeutic opportunities to target cell plasticity and disrupt tumor-stroma interactions[3,4]. Most adult cancers originate in chronically inflamed tissues due to exposure to intrinsic or extrinsic carcinogens. Inflammation is a hallmark of cancer and plays a critical role in tumor initiation, progression, and cell plasticity[3,4]. During tissue repair, the inflamed microenvironment undergoes extensive remodeling, altering both the quality and quantity of signaling molecules secreted by niche and stromal cells, including immune cells. Similar changes occur in TME[15]. Evidence linking inflammation and lineage plasticity has been observed in lung, breast, intestinal, pancreatic, prostate cancers[21,23,26,36,70,71]. In the pancreas, DCLK1+ progenitor cells represent quiescent acinar cells that aid in tissue repair. While mutant RAS expression in DCLK1+ cells alone is insufficient to induce cancer, experimentally induced pancreatitis results in acinar-to-ductal metaplasia, eventually leading to pancreatic ductal carcinoma[21,70,71]. For example, in chronic inflammatory bowel conditions, crosstalk between WNT and nuclear factor-kappa B pathways induces ectopic WNT signaling and β-catenin activation in differentiated intestinal epithelial cells. The resulting field defect leads to aberrant crypt expansion and activation of reprogramming machinery, predisposing epithelial cells to tumor initiation[4,36] (Table 1).

Similarly, CSCs exhibit remarkable ability to interact with its environment to exhibit plasticity. CSCs can transdifferentiate into vascular endothelial cells, pericytes, and other stromal components. This suggests that CSCs contribute to tumor growth, angiogenesis, and metastasis through lineage reprogramming[72]. Cell plasticity is well demonstrated in EMT and its reversal, MET. EMT enables tumor cells to lose polarity, acquire mesenchymal features, and penetrate the basement membrane - a critical step in invasion and metastasis. MET occurs when disseminated tumor cells reacquire epithelial characteristics, facilitating colonization at distant metastatic sites[14,73,74]. EMT is orchestrated by the TME, various signaling pathways, and microRNAs. Key transcription factors driving EMT include Snail, Twist, and Zeb-1, leading to loss of epithelial markers (e.g., E-cadherin, cytokeratin) and gain of mesenchymal markers (e.g., vimentin, N-cadherin). The presence of EMT markers in tumors correlates with invasion, metastasis, and poor prognosis[14].

Several other mechanisms exist, by which TME and associated chronic inflammation play pivotal roles in metastatic progression[73-75]. Tumor cells secrete chemokines, including C-X-C motif ligand 12 and osteopontin, which recruit bone marrow-derived stem cells into the tumor stroma[72,74]. Bone marrow-derived stem cells further facilitate recruitment of tumor-associated macrophages, cancer-associated fibroblasts, platelets, endothelial cells and other immune cells. Over the past decade, our understanding of cancer metabolism has also expanded significantly. Cells once considered passive bystanders in the TME such as cancer-associated fibroblasts, immune cells, and inflammatory cells are now recognized as active participants in metabolic remodeling. For instance, oxidative cancer cells utilize lactate as an energy source within the TME, while pancreatic cancer cells rely on alanine secreted by pancreatic stellate cells in the stroma. The transport of metabolites within tumor tissues, along with interactions between cancer cells and surrounding stromal cells, cooperatively reshape cancer metabolism, underscoring the complexity of metabolic regulation within the TME[69].

Once recruited, the stromal cells enhance tumor growth and survival by releasing chemokines, inflammatory markers, and EMT-inducing factors[4,74-77]. These factors are also responsible for pro-angiogenic signaling and homing of the distant tissue such as liver and lung prior to metastases. Systemic instigation, a phenomenon where TME-derived factors prime distant organs (e.g., lung, liver) for future metastases is an emerging concept in cancer progression[74].

Inflammation-driven changes in the extracellular matrix also significantly influence tumor progression. Increased collagen deposition and matrix stiffness are observed in breast and pancreatic tumors, correlating with poor survival[75-77]. High MET (hepatocyte growth factor receptor) expression and MET gene amplification in the stroma are linked to resistance against RAF inhibitors in lung cancer and melanoma[75]. Increased extracellular matrix stiffness promote EMT, invasion, metastasis via the TWIST1-G3BP2 pathway[3,76]. Lineage-tracing studies in pancreatic cancer mouse models demonstrate that EMT signatures emerge in premalignant lesions - even before clinical cancer diagnosis[77]. EMT transcription factors expressed in tumors resemble those involved in embryonic development and wound healing, reinforcing the concept that tumors modify developmental pathways for invasion and metastasis[78].

PLASTICITY AND DRUG RESISTANCE

Despite initial success with chemotherapy, radiation, and targeted therapies, many cancers exhibit residual tumor clones, leading to treatment resistance and metastatic recurrence. These resistant cancers often assume aggressive phenotypes with molecular features of stemness, dedifferentiation, or trans-differentiation, highlighting the role of cell plasticity in therapy resistance[50,53,79,80]. Understanding these plasticity-driven resistance mechanisms is crucial for developing effective therapeutic strategies that target both cancer cell plasticity and tumor heterogeneity, preventing disease progression and recurrence.

The differentiation state of a tumor significantly influences its therapeutic response. Cancer cells exploit reprogramming capabilities to survive treatment by overexpressing EMT transcription factors (Snail, Slug, Twist, Zeb) to acquire a hybrid epithelial-mesenchymal state, increasing stemness and tumor heterogeneity, which enhances chemotherapy, radiation, and immunotherapy resistance[79,80]. Other mechanisms include: Developing secondary mutations and metabolic reprogramming in CSCs[81-85], dysregulating ATP-binding cassette transporters, which mediate drug efflux and reduce intracellular drug accumulation[82] and adopting a mesenchymal phenotype, which is more therapy-resistant than epithelial-like tumors, as demonstrated in lung, pancreatic, and melanoma tumors[26,83-87].

Tumor reprogramming enables cancer cells to evade therapy-induced cell death, leading to cell reprogramming and emergence of therapy-resistant clones that sustain tumor growth and metastasis[83]. A striking example of tumor cell reprogramming is observed in patients with NSCLC treated with anti-EGFR therapy. Some EGFR-mutant cancers develop resistance by undergoing neuroendocrinal phenotypic switch to small cell lung cancer[84,85]. Histone acetylation can drive this trans-differentiation into a neuroendocrine phenotype. Similarly, CRPC demonstrates trans-differentiation, where adenocarcinoma of the prostate treated with AR inhibitors evolves into neuroendocrine prostate carcinoma, an aggressive and therapy-resistant subtype[42,88,89]. Histone acetylation at ASCL1 and SOX2 loci may contribute to this lineage switching.

Neuroendocrine gene signature is associated with metastasis and poor outcomes. The mechanisms underlying neuroendocrine differentiation in EGFR-treated NSCLC and AR-inhibited CRPC remains uncertain. However, lineage tracing studies suggest the combined loss of tumor suppressor genes (RB, p53, PTEN) in cancer cells, under the pressure of targeted therapy is necessary for the trans-differentiation switch[42,88-90]. Epithelial-to-neuroendocrine lineage plasticity frequently manifests following therapies targeted against driver oncogenes: Anti-EGFR therapy in EGFR-mutant lung adenocarcinoma, ALK inhibition in ALK-rearranged lung cancer, and anti-AR pathway therapy in prostate adenocarcinoma[91]. Loss of tumor suppressor genes may reactivate developmental homeobox transcription factors (Oct4, Sox4, Nanog), promoting a stem cell-like state and enabling therapy escape mechanisms[6,33,45]. This cell reprogramming allows tumors to bypass AR- or EGFR-dependence, conferring clonal growth advantages and promoting neuroendocrine differentiation[42,84,88,92]. Histone acetylation also plays a crucial role in promoting cancer cell plasticity by modifying chromatin accessibility and regulating gene expression as described earlier[52].

FUTURE DIRECTIONS

One of the greatest challenges with chemotherapy is the ability to predict clinical response and resistance at an individual level. Cancer cells can modify their genetic landscape, acquiring driver and passenger mutations, leading to tumor heterogeneity and phenotypic plasticity, as observed in lung and prostate cancers[42,84,88,92,93]. Currently, treatment decisions are based on mutational signatures and biomarkers to guide chemotherapy and targeted therapies. However, a major challenge lies in pre-emptively predicting resistance mechanisms, which remains an unmet clinical need. While large-scale cancer genomic studies have provided unprecedented insights into tumor biology, integrating these complex datasets into clinically actionable strategies remains difficult[94,95].

Advancements in artificial intelligence (AI), particularly deep neural networks (DNNs), offer promising solutions for modeling biological complexities and predicting therapeutic response[96]. DNNs, inspired by biological neural systems, excel at pattern recognition and data abstraction, making them valuable tools in oncology. DNN applications in cancer research include: Radiographic imaging and histopathology analysis[97,98], pharmacogenomics linking tumor genotypes to cellular phenotypes to identify precision therapy options[95-100]. Other application includes drug discovery and development, optimizing therapeutic strategies based on genomic data[101] and in synthetic biology for designing artificial biological pathways to target resistant cancer cells[102].

Recent pharmacogenomic studies have shown the importance of tissue lineage in predicting tumor response, sensitivity, and resistance. However, cancer’s dynamic reprogramming capability complicates treatment selection. Future advancements will likely integrate: Next-generation sequencing and single-cell RNA sequencing, providing high-resolution insights into tumor evolution and AI-driven deep learning models, capable of predicting cancer progression trajectories and guiding treatment decisions[102-105].

Google AlphaFold 2 (part of Alphabet Inc.’s AI research division, AF2) technology has significantly advanced understanding of protein structures, which is crucial for cancer research and drug development. By accurately predicting the three-dimensional (3D) structures of proteins, AF2 aids in identifying how mutations lead to cancer and assists in designing targeted therapies. By mapping out these structures, researchers can gain new insights into protein function, interaction pathways, and identify potential therapeutic targets that were previously hard to visualize[106-108]. It has the potential to significantly accelerate the discovery of new chemotherapy molecules by leveraging its ability to predict protein structures with remarkable accuracy. AF2 can model oncogenic mutations that alter protein stability and function and predict how mutations in these proteins might alter their structure and function, which is critical for understanding the mechanisms of resistance. By predicting protein 3D structures and identifying both active and potential allosteric binding sites, AF2 can also accelerate the development of new cancer drugs and therapies. For example, allosteric drugs, bind to the allosteric rather than the active sites, can induce conformational changes in proteins, affecting their activities. This can then alter the conformation of an active site that a drug-resistant mutation has created, permitting a blocked orthosteric drug to bind and this enables the design of more effective drugs that can synergize with traditional orthosteric drugs to bind and regain its efficacy[109].

AF2 may also assist with protein evolution and target prediction. Target prediction is for identifying novel drug targets and evaluating selectivity of drugs specially in cancers as they mutate and spontaneously on in response to chemotherapy. Computer-aided target prediction may help narrow the scope of target identification, which is often based on protein-ligand docking, usually called inverse docking however it requires knowledge of 3D structures of all possible protein targets. The AF2 structures provide opportunity to study and develop feasible target prediction methods. As cancer cells evolve and gains higher degree of complexity, their protein dynamics evolve. As such interesting properties may emerge such as larger gyration radii, higher coil fractions as well as slower vibrations. These observations of a relationship between organism evolution and protein evolution based on the structures of proteomes from 48 organisms was predicted by AF2. These findings may occur with tumor evolution too, which explains trans-differentiation often noted with advanced cancer under the pressure of treatments but is yet to be confirmed[110,111]. By studying and application of these principles’ researchers can design new drugs or molecules that can overcome resistance, either by binding to the mutated protein more effectively or by targeting alternative proteins involved in the resistance pathways.

CONCLUSION

The idea of deregulated cell plasticity being considered a hallmark of cancer emphasizes the adaptive nature of tumors, their ability to evade homeostatic regulation, and their capacity for resilience in the face of therapeutic intervention. Understanding the mechanisms that drive this plasticity could ultimately lead to the development of more effective therapies that target these adaptive processes and prevent the tumor from exploiting its ability to continuously evolve and resist treatment. Google’s AlphaFold represents a game-changer in chemotherapy drug discovery by providing a fast and accurate method to predict protein structures, enabling researchers to identify new drug targets, design better molecules, and understand the complex biology of cancer at a level of detail that was previously unattainable. By integrating next-generation sequencing and single-cell RNA sequencing, AI-driven deep learning models, and AlphaFold’s predictions with experimental data and existing drug discovery techniques, we could see faster development of more effective, specific, chemotherapy agents in the future. By merging these technologies, precise, durable, and truly personalized cancer therapy could become a clinical reality, revolutionizing the way we approach cancer treatment.