The progression of tumors from localized growth to metastatic malignancy is governed by a complex array of interacting signaling pathways and molecular mechanisms. As discussed in the previous section on evolutionary models, tumor evolution is a dynamic, multifactorial process shaped by both intrinsic and extrinsic factors. The evolutionary trajectory of cancer cells is influenced not only by genetic alterations inherent to the tumor but also by the cellular and environmental contexts in which these alterations occur. These regulatory mechanisms operate across three critical dimensions—intracellular, extracellular, and exogenous levels—each playing a crucial role in orchestrating tumor evolution (Fig. 3).
Fig. 3
Overview of multidimensional regulatory factors in tumor evolution. Tumor evolution is regulated by multiple driving factors from three dimensions: cell-intrinsic level (including genomic instability, epigenetic plasticity and metabolic reprogramming), cell-extrinsic level (including tumor microenvironment, immune escape and intercellular communication), and exogenous level (including therapeutic interventions, environmental factors, characteristics of patients). These factors interact closely and ultimately manifest in the form of tumor heterogeneity in both temporal and spatial dimensions. Abbreviations: MSI, microsatellite instability; CIN, chromosomal instability; EMT, epithelial–mesenchymal transition
At the cellular level, intrinsic factors such as genomic instability, epigenetic modifications, and metabolic reprogramming drive the clonal diversity that characterizes cancer progression. These factors, often acting synergistically, facilitate the selection of tumor cells with growth advantages, enabling them to survive and proliferate under both physiological and therapeutic stresses. Meanwhile, the tumor microenvironment (TME), consisting of stromal cells, immune cells, and extracellular matrix components, exerts a significant influence on tumor evolution, promoting processes such as invasion, immune evasion, and resistance to treatment. Additionally, extrinsic factors, including environmental influences and therapeutic interventions, introduce selective pressures that further shape the tumor’s evolutionary landscape.
As discussed below, we will examine the signaling pathways and molecular mechanisms that regulate tumor evolution across these three dimensions. By integrating insights from these diverse mechanisms, we aim to develop a comprehensive understanding of tumor evolution, ultimately facilitating the creation of more effective and targeted therapeutic strategies.
Cell-intrinsic regulatory mechanisms in tumor evolution
Cell-intrinsic regulatory mechanisms are central to understanding how cancer cells acquire and sustain their malignant phenotypes. These intrinsic factors, such as genomic instability, epigenetic plasticity, and metabolic reprogramming, are pivotal in enabling tumor cells to adapt to selective pressures, accumulate mutations, and proliferate uncontrollably. A deeper understanding of the molecular pathways and cellular processes governing these mechanisms is crucial for elucidating the evolution of tumors and their resistance to therapeutic interventions.
Genomic instability
Genomic instability is one of the most defining features of tumor evolution. It manifests in various forms, such as genetic mutations, chromosomal aberrations, and aneuploidy. These changes can drive the acquisition of oncogenic traits, clonal diversity and resistance to therapies. A thorough understanding of the molecular mechanisms driving genomic instability is crucial for understanding how tumors evolve and adapt to both endogenous and exogenous stresses.
One of the most significant contributors to gene mutations in various cancer types is the activity of the APOBEC3 family of cytidine deaminases. APOBEC3 enzymes, particularly APOBEC3A and APOBEC3B, induce mutations through the deamination of cytosine residues, leading to “C to T” and “C to G” substitutions. The mutational signatures associated with APOBEC3 activity have been identified as a hallmark of multiple cancers, contributing to tumor mutational burden (TMB) and overall genomic instability.162,163 APOBEC3-mediated mutations are particularly prevalent in late-stage tumor subclones and contribute to the evolution of tumor heterogeneity. Studies have shown that APOBEC3-induced mutations play a key role in the branching evolution of tumors, contributing to clonal diversification and resistance to therapeutic interventions.164,165 For instance, in bladder cancer, APOBEC3G expression was found to induce a distinct mutational signature that correlated with poor patient survival.163 This mutational process is not uniform but rather episodic, with bursts of mutagenesis triggered by factors such as viral infection, DNA replication stress, and inflammation.166 Moreover, APOBEC3 activity is not limited to driving somatic mutations. It also plays a role in modulating the immune environment of tumors. While some cancers, such as melanoma and breast cancer, exhibit immune-activating properties due to APOBEC3-induced mutations, others, such as renal cell carcinoma, exploit APOBEC3-induced mutations to escape immune surveillance.167,168,169,170 These dual roles highlight the complex nature of APOBEC3 in tumor progression, offering potential therapeutic targets to control its mutagenic activity.
Another manifestation of genomic instability in tumors is the formation of extrachromosomal DNA (ecDNA), which plays a critical role in tumor evolution. EcDNA is a circular, acentric DNA molecule distinct from chromosomal DNA. These DNA fragments, typically ranging from 50 kb to 5 Mb in size, carry oncogenes, immunoregulatory genes, and enhancers, playing a crucial role in tumor progression by facilitating high-level transcription and promoting cancer heterogeneity.171,172,173 The unique nature of ecDNA, particularly its lack of centromeres, enables it to be distributed unequally to daughter cells during cell division, resulting in heterogeneous tumor populations and promoting clonal diversity.174,175 This irregular inheritance creates a dynamic environment within the tumor, where subclonal populations emerge rapidly and adapt to selective pressures, accelerating tumor progression.176,177 EcDNA can also undergo structural rearrangements, contributing to further genomic instability.167 These rearrangements enable ecDNA to amplify oncogenes and facilitate rapid evolutionary changes. Multiple ecDNA species may coexist within a single tumor, and their structural complexity tends to increase over time, particularly as the tumor progresses.167,178,179 The presence of different ecDNA species contributes to the formation of a multifocal and highly heterogeneous tumor population, fostering adaptability and enhancing the likelihood of therapy resistance.171,177 Moreover, these extrachromosomal elements can reintegrate into chromosomal DNA, leading to genetic rearrangements that influence gene activation and contribute to the malignant transformation of tumor cells.180
The role of ecDNA in tumor evolution is further underscored by its interaction with mutagenic processes. More than 30% of ecDNAs are associated with APOBEC3-induced kataegis, a mutagenic process that contributes to the sequence diversity of ecDNA and drives the evolution of therapy-resistant clones.181,182 As mentioned before, these APOBEC3-mediated mutations are particularly significant in promoting genomic instability, enhancing the diversity of ecDNA sequences, and contributing to the adaptability of cancer cells under therapeutic stress.181,183 The ability of ecDNA to accumulate mutations, coupled with its rapid inheritance and amplification, makes it a central player in the evolution of malignancies and their resistance to treatment.184
Furthermore, ecDNA can form transcriptional hubs, where clusters of ecDNA molecules aggregate in specific regions of the nucleus.185 These hubs are associated with increased transcriptional activity and play a crucial role in regulating gene expression within the tumor.186 Transcriptional regulation within ecDNA hubs is closely tied to proteins such as BRD4, which mediate the formation of these hubs and facilitate the transcription of oncogenes that drive tumor progression.185 Therefore, the dynamic properties of ecDNA, including its ability to undergo structural rearrangements, amplify oncogenes, and interact with mutagenic pathways such as APOBEC3, highlight its critical role in promoting cancer heterogeneity and accelerating tumor evolution. This phenomenon is particularly prevalent in cancers exhibiting high genomic instability, such as those induced by CRISPR/Cas9-based mutagenesis. The amplification of ecDNA can increase oncogene dosage and enhancer-driven transcription, thereby supporting intratumoral heterogeneity (ITH) and providing a reservoir for further mutational evolution.187 Notably, the formation of ecDNA has been associated with resistance to targeted therapies and contributes to the adaptability of tumors in response to environmental stressors and therapeutic pressures.
The role of genomic instability in tumor evolution is not restricted to mutagenic enzymes and ecDNA formation. Chromosomal instability (CIN) is also a critical driver of tumor evolution. CIN is primarily caused by errors during the mitotic process, leading to chromosomal structural and numerical abnormalities, including copy number alterations (CNAs), aneuploidy, chromoplexy and chromothripsis. Recent studies have demonstrated that CIN enhances tumor evolution by providing rapid adaptive opportunities through genome doubling and copy number alterations, promoting tumor aggressiveness and metastasis. In a comprehensive study of lung adenocarcinoma (LUAD) from the TRACERx cohort, the role of CIN in shaping ITH was emphasized.188 Tumors with high levels of CIN exhibited increased chromosomal complexity and a higher frequency of subclonal copy number variations, which were linked to phenotypic changes such as increased proliferation and reduced clonality. These findings underscore the pivotal role of CIN in the development of tumor heterogeneity, which is essential for the clonal evolution of aggressive tumor subpopulations.
Aneuploidy, the presence of an abnormal number of chromosomes, is another fundamental aspect of genomic instability in cancer. It often results from CIN and plays a crucial role in facilitating tumor evolution by allowing cells to adapt to selective pressures, including therapeutic interventions. High levels of aneuploidy have been associated with poor patient outcomes due to their role in promoting tumor progression and treatment resistance. The mechanisms by which aneuploidy drives tumor evolution include the alteration of gene dosage, leading to the upregulation of oncogenes or the loss of tumor suppressor genes.189 This can result in the activation of key signaling pathways that promote cell survival, proliferation, and metastasis. Aneuploidy has been described as a “bet-hedging” mechanism in tumor evolution, where genetic instability provides an opportunity for cells to rapidly adapt to changing environmental conditions, including therapeutic pressures.190 By accumulating diverse chromosomal aberrations, tumor cells enhance their adaptability, allowing them to thrive under selective pressures that would typically lead to cell death. This ability to tolerate high levels of CIN and aneuploidy is particularly relevant in the context of cancer treatment, where treatment resistance is a major barrier to therapeutic success. A recent study examining colorectal cancer organoids demonstrated the role of CIN in promoting tumor progression, with cells exhibiting varying levels of tolerance to mitotic errors.191 This variability in CIN tolerance led to the formation of distinct karyotypic landscapes, contributing to tumor heterogeneity and resistance to treatment. Such findings highlight the dynamic nature of CIN and its critical role in shaping the tumor genome and its response to therapy.
Chromoplexy and chromothripsis are representative examples of single catastrophic events that can drive tumor evolution, suggesting their involvement in the punctuated model of cancer evolution. Rather than a gradual accumulation of mutations, chromoplexy and chromothripsis can cause sudden and dramatic changes in the genome, leading to rapid evolutionary shifts in tumor populations. This model contrasts with traditional linear or branching models of tumor evolution and highlights the importance of catastrophic genomic events in shaping tumor heterogeneity and therapeutic resistance.
Chromoplexy refers to a catastrophic genomic event involving complex, simultaneous chromosomal rearrangements.192 Chromoplexy is triggered by the occurrence of multiple double-strand breaks in the genome, typically in regions of the genome prone to such damage. These breaks lead to highly complex repair mechanisms, often involving nonhomologous end joining (NHEJ) or alternative end-joining pathways, which result in the rearrangement of chromosomal segments.193,194,195 Disrupted chromosomal regions can lead to the formation of fusion genes, creating new oncogenic drivers that promote tumor progression. In prostate cancer, for example, chromoplexy often involves the rearrangement of genes such as the fusion between TMPRSS2 and ERG, contributing to cancer cell proliferation and survival.196 These fusions serve as a significant driver of tumorigenesis, highlighting the role of chromoplexy in creating oncogenic gene combinations. Moreover, several key signaling pathways are associated with chromoplexy-driven tumor evolution. For example, the androgen receptor (AR) signaling pathway in prostate cancer is influenced by chromoplexy-induced gene fusions, such as TMPRSS2:ERG.192 These fusions activate downstream targets involved in cell survival and proliferation. Additionally, chromoplexy can affect pathways such as PI3K/AKT and MAPK signaling, which are crucial for cancer cell growth and survival. Chromoplexy-induced mutations contribute to the evolution of aggressive cancer phenotypes, as the genomic chaos generated by chromoplexy can alter the function of key regulatory genes, facilitating tumor survival and metastasis. This phenomenon underscores the role of chromoplexy in the early stages of cancer development, where a single catastrophic event can lead to a cascade of mutations and clonal evolution.
Chromothripsis refers to a catastrophic genomic event characterized by massive, localized chromosomal fragmentation followed by incomplete repair, leading to highly complex genomic rearrangements.197 Unlike chromoplexy, which involves multiple chromosomes, chromothripsis typically affects a single or few chromosomes, generating a fragmented, shattering pattern.159 These fragmented chromosomes are then improperly repaired, resulting in structural variations such as deletions, duplications, translocations, and inversions, all of which contribute to genomic instability in tumors. Chromothripsis often occurs during cellular stress, such as replication or telomere dysfunction, and has been linked to abnormal mitosis. This leads to chromosomal shattering, which, when repaired by NHEJ or alternative end joining (alt-EJ), results in an array of structural variants. These variants disrupt the normal function of key tumor suppressor genes and oncogenes, thus fostering the clonal evolution of the tumor. In prostate cancer, for example, chromothripsis contributes to the formation of fusion genes such as TMPRSS2:ERG, which further drives tumor progression. Additionally, in breast cancer, chromothripsis has been observed in a significant proportion of cases, particularly in metastatic tumors.197 This event has been linked to the generation of recurrent fusions that drive tumor progression, with several well-known oncogenes, such as CCND1, ERBB2, and CDK12, being affected. Chromothripsis also influences the activation of pathways such as the PI3K/AKT and MAPK pathways, which promote cell survival, growth, and metastasis. Interestingly, chromothripsis is frequently associated with kataegis, a mutational signature characterized by clustered point mutations, suggesting that chromothripsis plays a dual role in both structural and mutational changes that facilitate tumor evolution. Strategies to target CIN, such as inhibitors of mitotic checkpoint kinases or drugs that induce chromosomal instability, may offer new avenues for cancer therapy. In addition, recent advancements in genomic sequencing technologies, such as single-cell sequencing and spatial transcriptomics, have allowed for a more detailed characterization of CIN and aneuploidy at the single-cell level. These technologies provide a deeper understanding of how chromosomal aberrations contribute to tumor evolution and offer the potential for more personalized treatment strategies.
Furthermore, research involving single-cell copy number profiling and exome sequencing from the same breast cancer patients has revealed that CNAs emerge early in evolution as punctuated bursts, followed by stable expansion. In contrast, point mutations accumulate gradually throughout the tumor’s progression, driving clonal expansions over time.121 Similarly, studies employing whole-genome sequencing of matched longitudinal breast tumor samples from hyperplasias, ductal carcinoma in situ (DCIS), and invasive carcinomas have demonstrated that copy number evolution occurs early, while point mutations evolve more gradually and branch throughout disease progression.198 This indicates the correlation between tumor evolution models and the molecular mechanisms driving tumor evolution. In the early stages of tumor development, genomic instability events, such as CNAs, chromoplexy, and chromothripsis, emerge rapidly in an explosive manner. These events drive tumor evolution in accordance with the punctuated equilibrium model. In contrast, point mutations, which accumulate gradually due to defects in DNA repair pathways and errors in DNA polymerases, often drive branching tumor evolution.199,200
In conclusion, genomic instability is a central driver of tumor evolution, contributing to the accumulation of mutations, the formation of clonal diversity, and the development of resistance to therapies. The mechanisms underlying genomic instability—ranging from the activity of mutagenic enzymes such as APOBEC3G to the formation of ecDNA and the accumulation of chromosomal aberrations—provide critical insights into how tumors evolve and adapt. Understanding these processes is essential for identifying potential therapeutic targets aimed at disrupting the evolutionary dynamics of tumors and preventing the emergence of therapy-resistant clones.
Epigenetic plasticity
Epigenetic modifications play a crucial and dynamic role in tumor evolution, contributing to the establishment of cellular diversity and the adaptability of cancer cells within the tumor microenvironment (TME). Unlike genetic mutations, which are typically irreversible, epigenetic changes are reversible, allowing for plasticity that enables tumor cells to respond to environmental pressures such as hypoxia, immune attack, and therapeutic intervention. These alterations, which include DNA methylation, histone modification, and chromatin remodeling, provide a mechanism for tumor cells to acquire malignant phenotypes without changing the underlying genetic sequence. This ability to undergo epigenetic reprogramming is a key feature that drives tumor progression, metastasis, and therapeutic resistance. In this section, we will discuss how epigenetic plasticity drives tumor evolution, focusing on various molecular mechanisms and pathways.
DNA methylation is a critical epigenetic modification that contributes to the regulation of gene expression in tumors. Aberrant DNA methylation patterns, such as hypermethylation of tumor suppressor genes and hypomethylation of oncogenes, are frequently observed in various cancers and contribute to tumorigenesis and progression. Early in tumorigenesis, abnormal DNA methylation can establish a foundation for subsequent genetic alterations, thereby contributing to tumor heterogeneity and clonal evolution. These methylation changes are not random but often follow specific patterns that can provide insight into tumor evolution.201
In the context of tumor progression, DNA methylation plays a pivotal role in regulating the expression of genes involved in immune evasion. For example, the expression of immune checkpoint molecules such as PD-L1 can be modulated by methylation at gene promoter regions, allowing tumors to escape immune surveillance. This process is further compounded by the silencing of immune-related genes, which can be reversed by demethylating agents, offering potential therapeutic strategies to counteract immune evasion.202 As tumors evolve, the accumulation of DNA methylation changes can also lead to the stabilization of specific gene expression profiles, which support the adaptive capabilities of the tumor under therapeutic pressures.
In addition to DNA methylation, histone modifications play a critical role in regulating gene expression during tumor evolution. Modifications such as acetylation, methylation, phosphorylation, and ubiquitination of histones can alter the chromatin structure, either promoting or repressing transcriptional activity. Histone acetylation, for example, is associated with an open chromatin structure and active transcription, while histone methylation can result in either gene activation or silencing depending on the specific context and the histone residues involved. One of the most well-studied histone modifications in cancer is H3K27ac, which marks active enhancer regions in the genome. It has been demonstrated that the activation of enhancers, as indicated by increased H3K27ac levels, is associated with phenotypic heterogeneity within tumors.203 These changes in enhancer activity reflect the dynamic adaptation of tumors to their environment and are closely linked to therapeutic resistance. In the case of endocrine-resistant breast cancer, for example, changes in enhancer activity drive the expansion of specific subclones that exhibit resistance to treatment, highlighting the role of epigenetic regulation in modulating tumor heterogeneity.203
Furthermore, chromatin remodeling complexes such as SWI/SNF and the chromatin-modifying enzymes EZH2 and KDM1A are frequently altered in cancer.204,205,206 These proteins regulate chromatin accessibility and the expression of key oncogenes and tumor suppressors. Mutations or altered expression of these chromatin regulators can contribute to tumorigenesis by altering the transcriptional landscape, leading to uncontrolled cell proliferation and survival. These modifications often result in an epigenetic “memory” that persists even in the absence of the initial genetic alteration, providing a lasting advantage to the tumor.204
Epigenetic plasticity is not only critical for tumor initiation and progression but also plays a central role in metastasis and therapeutic resistance. Tumor cells that undergo epithelial-to-mesenchymal transition (EMT) represent a well-characterized example of phenotypic plasticity, while phenotypic plasticity can also manifest through other processes, including the acquisition of stem-like states and therapy-induced transcriptional reprogramming.207,208 EMT-associated epigenetic reprogramming allows tumor cells to acquire invasive and migratory capabilities. The dynamic and reversible nature of EMT, as mediated by epigenetic modifications, enables cancer cells to adapt to different microenvironmental niches and resist therapeutic intervention.209 This process is often driven by transcription factors such as ZEB1, which regulate both the expression of EMT markers and the epigenetic landscape of the cell.
Epigenetic reprogramming also contributes to the acquisition of therapy resistance. Cancer cells can undergo reversible epigenetic changes that allow them to survive in the face of chemotherapy or immunotherapy. For instance, certain subpopulations of tumor cells may enter a dormant or “persister” state through chromatin modifications, enabling them to evade the cytotoxic effects of treatment. Upon cessation of therapy or changes in the microenvironment, these cells can “wake up”, initiating tumor relapse and contributing to poor therapeutic outcomes. Such epigenetic mechanisms, which are independent of genetic mutations, underscore the complexity of cancer evolution and highlight the need for novel therapeutic strategies that target both genetic and epigenetic drivers of resistance.210
A significant aspect of epigenetic plasticity in tumor evolution is its role in immune evasion. Tumor cells frequently exploit epigenetic mechanisms to modulate the expression of immune-related genes, thereby escaping immune surveillance. The epigenetic regulation of genes involved in immune recognition, such as MHC molecules, plays a key role in determining whether tumor cells are visible to the immune system. The CSDE1 protein regulates immune recognition by modulating STAT1 phosphorylation.211 The differential expression of CSDE1 in nascent tumorigenic cells, regulated by SMYD3-mediated H3K4 trimethylation, dictates whether these cells exhibit high or low immunogenicity. This epigenetic pathway represents a crucial determinant of tumor immune visibility and contributes to immune evasion during early tumor evolution. In particular, the modulation of CSDE1 expression through mechanical signaling in the TME underscores the importance of the microenvironment in shaping the epigenetic landscape of tumor cells and their ability to evade immune detection.211
Metabolic reprogramming
As tumors progress, they encounter environmental stresses such as hypoxia, nutrient deprivation, and immune surveillance, which trigger changes in metabolic pathways to sustain cellular functions and promote malignancy. These metabolic alterations offer a significant advantage in tumor evolution. The following discussion explores the molecular mechanisms and signaling pathways involved in metabolic reprogramming and their contribution to the dynamic evolution of tumors.
The Warburg effect, first described by Otto Warburg in the 1920s, is a fundamental aspect of metabolic reprogramming in cancer cells.212 This phenomenon is characterized by increased glucose consumption and lactate production, even in the presence of oxygen, a process known as aerobic glycolysis. Tumor cells often exhibit this metabolic shift, which supports rapid ATP generation and the biosynthesis of key macromolecules needed for growth and proliferation. This shift away from oxidative phosphorylation is thought to be driven by oncogenic signaling pathways, including the activation of the PI3K-AKT-mTOR and MYC pathways, which promote glycolysis by upregulating glycolytic enzymes such as hexokinase 2 (HK2) and lactate dehydrogenase A (LDHA).213 The Warburg effect is not just a metabolic quirk but an adaptive response to the microenvironment of rapidly growing tumors, which often face limited oxygen and nutrient availability. The metabolic flexibility conferred by the Warburg effect allows tumor cells to thrive in hypoxic conditions and provides a mechanism for survival in hostile environments, a key feature in tumor progression and metastasis.214
In addition to the Warburg effect, cancer cells undergo extensive reprogramming of multiple metabolic pathways to meet the energy demands of their rapid growth and survival in a hostile microenvironment. These changes include alterations in glucose metabolism, lipid metabolism, amino acid metabolism, and mitochondrial function. Glucose is a critical nutrient for cancer cells, and its metabolism is heavily altered in tumors. In glycolysis, glucose is converted to pyruvate, which, under normal conditions, enters the mitochondria for oxidative phosphorylation. However, in many cancers, pyruvate is converted to lactate by LDHA, even under aerobic conditions, ensuring continued ATP production via glycolysis.215,216 This shift is particularly evident in aggressive cancers and is a key feature of the metabolic adaptation that supports tumor progression. Furthermore, the activation of the PI3K-AKT-mTOR pathway enhances glucose uptake through increased expression of glucose transporters (GLUTs) and key glycolytic enzymes.217 This metabolic shift is essential for the rapid cell division and growth seen in tumors.218
Lipid metabolism is another critical aspect of metabolic reprogramming in tumor cells. Tumor cells increase fatty acid synthesis to provide lipids for the construction of cell membranes and to produce signaling molecules. Fatty acid synthase (FASN) is a key enzyme in this process, and its overexpression is common in many cancer types, including breast cancer.219 Elevated levels of FASN drive the synthesis of fatty acids, which are integral for the formation of phospholipids in cellular membranes, a requirement for tumor cell proliferation. Additionally, altered lipid metabolism contributes to the regulation of cell signaling pathways involved in tumor progression. For example, the accumulation of specific lipid species can activate oncogenic pathways such as the PI3K-AKT pathway, further fueling the aggressive growth of tumors.220
Furthermore, amino acids are essential for protein synthesis and cellular growth, and their metabolism is often reprogrammed in cancer cells. One prominent example is the reprogramming of glutamine metabolism. Glutamine is used as a carbon and nitrogen source for biosynthesis and mitochondrial energy production in cancer cells. In some tumors, the glutamine transporter SLC1A5 is upregulated to increase glutamine uptake, and glutamine is converted into α-ketoglutarate, which enters the TCA cycle to fuel anabolic pathways.221 Moreover, altered amino acid metabolism can influence the tumor microenvironment (TME) by modulating immune cell function. For instance, elevated levels of tryptophan metabolites, such as kynurenine, can suppress T-cell activation and promote immune evasion.222
As tumors evolve, they experience metabolic selection pressures that drive the acquisition of specific genetic alterations, particularly those that promote survival under metabolic stress. Copy number alterations (CNAs) in metabolic genes are common in tumors, particularly those involved in glycolysis, fatty acid synthesis, and amino acid metabolism.213,223 These alterations provide a survival advantage to tumor cells by enhancing their ability to adapt to changing metabolic conditions, thus driving tumor progression. For example, the amplification of glycolytic genes such as hexokinase and enolase is frequently observed in tumors with high metabolic activity, highlighting the selective advantage conferred by metabolic reprogramming.223
Metabolic reprogramming is a critical factor in tumor evolution, enabling cancer cells to adapt to the harsh conditions of the tumor microenvironment while promoting uncontrolled growth and survival. Key metabolic alterations, including the Warburg effect, changes in lipid and amino acid metabolism, and adaptations to hypoxic stress, are all crucial to understanding how tumors evolve over time. These metabolic shifts are not only essential for tumor growth and progression but also represent potential therapeutic targets for future cancer treatments. By targeting these metabolic pathways, it may be possible to disrupt tumor evolution and improve clinical outcomes for patients.
In summary, cell-intrinsic mechanisms are essential drivers of tumor evolution, shaping the genetic and epigenetic landscape of tumors while facilitating their adaptation to therapeutic pressures. Genomic instability, through mechanisms such as APOBEC3-mediated mutations and the formation of ecDNA, contributes to the diversification of tumor clones and the development of resistance to therapies. Epigenetic plasticity, marked by reversible changes in DNA methylation, histone modifications, and chromatin remodeling, further enhances tumor adaptability and progression, enabling cells to evade immune surveillance and therapeutic interventions. Additionally, metabolic reprogramming allows tumor cells to thrive under stress conditions, fostering uncontrolled growth and contributing to metastasis. These intrinsic mechanisms work synergistically to promote the heterogeneity and aggressiveness of tumors, underscoring the complexity of cancer evolution. As we transition to the next section, we will explore how external factors, such as the tumor microenvironment and immune system, also play a pivotal role in driving tumor progression and shaping evolutionary dynamics.
Cell-extrinsic regulatory mechanism of tumor evolution
The complexity of the tumor ecosystem is shaped in space and time by diverse cell-extrinsic factors stemming from the tumor microenvironment, immune system, intercellular interactions and other extracellular driving factors. These external cues do not act passively. Instead, they actively sculpt clonal dynamics, influencing which cancer cell populations survive, expand, or are eliminated.
Tumor microenvironment
While these extrinsic factors convey evolutionary pressure to the distinct clones within the tumor, cancer cells themselves are also capable of reconstructing a tumor-supportive microenvironment by recruiting and reprogramming noncancerous host cells and by remodeling angiogenesis and the extracellular matrix (ECM).224,225 Hypoxia in the tumor microenvironment induces vascular endothelial growth factor (VEGF) expression and polarizes tumor-associated stromal cells (e.g., tumor-associated macrophages, tumor-associated fibroblasts), driving tumor microenvironmental remodeling, which exacerbates intratumoral heterogeneity and evolutionary potential.226 In addition to its effects on stromal dynamics, hypoxia also promotes genomic instability through multiple mechanisms, including the accumulation of reactive oxygen species (ROS)-mediated DNA damage, impaired replication restart fidelity, and suppression of DNA damage response (DDR) pathways, including mismatch repair and methylation silencing of BRCA1.227
Cancer cells actively remodel their surroundings through the recruitment and reprogramming of nonmalignant host cells, including fibroblasts, immune cells, and endothelial cells. Cancer-associated fibroblasts (CAFs), for instance, secrete a wide array of mitogenic growth factors, including hepatocyte growth factor (HGF), fibroblast growth factors (FGFs), insulin-like growth factor-1 (IGF-1), and stromal-derived factor 1 (SDF-1/CXCL12).228,229,230 These paracrine signals complement oncogenic mutations and enable clonal expansion even under therapeutic selective pressure. Moreover, immunosuppressive factors in the TME, such as CXCL12, CXCL14, and IL-6, secreted by inflammatory-like cancer-associated fibroblasts pan-iCAF and pan-iCAF-2, contribute to tumor evolution by promoting immune evasion, cancer stem cell renewal, and the polarization of tumor-associated macrophages.231
Immune system
The immune system exerts one of the most potent forms of selection on developing tumors. The immunoediting framework—comprising elimination, equilibrium, and escape—describes how nascent tumors are initially constrained by immune surveillance but may eventually adapt to survive in an immunosuppressive niche.232 Immune escape is also an important driver of tumor evolution. For example, the progression of breast ductal carcinoma in situ (DCIS) to invasive ductal carcinomas was accompanied by a transition to a suppressed immune milieu characterized by fewer activated CD8 + T cells, increased PD-L1 and CTLA4 expression, more regulatory T cells (Tregs), and less diverse T-cell receptor (TCR) clonotypes.233 Weeden et al. revealed that tissue-resident memory T (TRM) cells, while normally providing local immune defense, can shape tumor evolution through chronic immune pressure.234 In ever-smokers, heightened TRM-like activity prior to tumor onset enhances immune surveillance but also promotes immune escape by selecting tumor clones that downregulate MHC class I and resist immune checkpoint blockade (ICB).
Importantly, immune escape should not be viewed as a static or terminal event but rather as a dynamic and evolving process shaped by selective pressures throughout tumor progression. Tumor clones may acquire partial escape mechanisms early on and refine or diversify these strategies as immunologic conditions change, particularly under the influence of therapy. As a result, the efficacy of immunotherapeutic interventions—such as immune checkpoint blockade—depends not only on the presence of immune-competent cells but also on the specific timing, mechanism and evolutionary stage of immune escape.
To address this challenge, real-time evolutionary monitoring of tumor-immune dynamics is essential. Approaches such as circulating tumor DNA (ctDNA) profiling, neoantigen landscape tracking, and spatial immunogenomic analysis can help identify emerging escape subclones and adapt treatment strategies accordingly. By integrating such evolutionary surveillance into clinical decision-making, it may be possible to anticipate immune evasion, guide rational combinations, and prolong the window of therapeutic vulnerability.
Intercellular interactions
Recent advances in computational analysis and modeling have revealed a diversity of intercellular signaling networks, revealing how complex intercellular interactions are integrated to promote the evolution of tumor.225,235 There are multiple mechanisms by which this intercellular dialog is regulated, including cell‒cell contact and paracrine signaling.236 Contact-dependent communication is mediated by adhesion molecules, including integrins, cadherins, selectins, and immunoglobulin superfamily members, and via gap junctions and tunneling nanotubes.225 For instance, aberrant glycan sialylation on cancer cells regulates numerous interactions, including interactions with Siglec-expressing immune cells, promoting immune evasion and tumor evolution.237 In addition to direct cell‒cell contact, paracrine signaling through the release of cytokines, chemokines, growth factors, and proteases is also critical for intercellular communication.225 These molecules are secreted in response to cancer-intrinsic features and cellular stress, exerting direct and indirect actions on target cells by ECM remodeling or binding to receptors. Moreover, extracellular vesicles (EVs), particularly exosomes, have emerged as pivotal carriers of intercellular messages. These vesicles transport bioactive molecules—proteins, lipids, RNAs, and DNA fragments—to recipient cells, influencing their behavior in a context-dependent manner. For example, melanoma-derived exosomes enriched in PD-L1 can suppress T-cell activation systemically, contributing to immune evasion.238 In breast cancer, exosomal miR-105 derived from metastatic cells has been shown to disrupt tight junctions in endothelial cells by targeting ZO-1, promoting vascular permeability and premetastatic niche formation.239 Similarly, glioma-derived exosomes carrying EGFRvIII can transform neighboring astrocytes, facilitating an invasive tumor microenvironment.240
Spatial transcriptomics and tools such as CellPhoneDB241 and NicheNet242 have further uncovered complex cell–cell communication networks within tumors, revealing that stromal and immune cell subsets preferentially interact with specific cancer subclones. These interaction patterns are not static but evolve over time in response to selective pressures such as nutrient gradients or immune responses.243 Aberrant glycosylation patterns—such as hypersialylation—also mediate intercellular crosstalk by engaging inhibitory Siglec receptors on immune cells, promoting immune tolerance and escape.244
Other extracellular driving factors
In addition to well-characterized components such as the tumor microenvironment, immune system and intercellular interactions, biophysical and physiological factors within the tumor ecosystem also exert profound evolutionary influence.
Temporal fluctuations in blood flow are a hallmark of malignant tumors, yet their evolutionary significance has often been overlooked. These variations stem from the eco-evolutionary dynamics of tumor angiogenesis, where cancer cells—acting as local units of selection—respond solely to proximal microenvironmental cues.245 As a result, neovessels commonly emerge from the nearest available vasculature without coordination or regard for functional capacity, giving rise to a chaotic vascular architecture and pronounced spatial and temporal heterogeneity in perfusion and tissue conditions.246
This unstable hemodynamic landscape imposes shifting selective pressures that fuel cancer cell adaptation, promoting phenotypic plasticity, enhanced motility, invasiveness, and metabolic versatility.247 Clinically, fluctuating blood flow intensifies intratumoral hypoxia, hampers effective drug delivery, and accelerates the selection of resistant clones—collectively shaping treatment outcomes and long-term disease evolution.248
Beyond fluctuations in blood flow and perfusion, a diverse array of biophysical forces—including mechanical stress, interstitial pressure, osmotic gradients, and extracellular acidosis—further contribute to tumor evolution by shaping the selection landscape. As tumors grow within confined anatomical spaces, cancer cells are subjected to increasing solid stress and tensile strain from the surrounding stroma and ECM. These forces influence cell fate decisions by modulating mechanotransduction pathways, such as integrin–FAK–YAP/TAZ signaling, which in turn promote survival, proliferation, and stem-like phenotypes in subclones capable of withstanding mechanical challenge.247 Chemical and ionic gradients also impose powerful selective constraints. Acidic extracellular pH, resulting from aerobic glycolysis and impaired clearance of metabolic waste, not only favors clones with altered pH-regulating transporters (e.g., CAIX, MCT4) but also induces mutagenic stress and immune suppression.249 Subclones that thrive in acidic environments often exhibit enhanced treatment resistance and invasive potential.250 Hyperosmotic stress, on the other hand, as seen in densely packed or poorly perfused regions, can activate stress-adaptive signaling such as the p38 MAPK and HSP pathways, selecting for more resilient cellular phenotypes.251
Collectively, these physical and chemical factors define microenvironmental niches that filter cancer cell populations based on mechanical resilience, metabolic adaptability, and signaling plasticity. The result is a highly heterogeneous tumor composed of subclones fine-tuned to distinct ecological constraints—an evolutionary mosaic continually reshaped by nongenetic, extrinsic forces.
Exogenous regulatory mechanism of tumor evolution
Building upon the cellular-extrinsic factors discussed in the previous section, tumor evolution is also profoundly influenced by a range of exogenous regulatory mechanisms. These factors, which include therapeutic interventions, environmental influences, and the diverse clinical characteristics of the patient population, serve as additional evolutionary pressures that shape the dynamics of tumor progression. Unlike intrinsic cellular alterations, exogenous factors exert a force on tumors from outside the cellular realm, providing selective pressures that influence clonal selection, adaptation, and therapeutic resistance. While these external forces may act synergistically with intrinsic mutations and cellular plasticity, their contribution to the tumor evolutionary process often determines the trajectory of the disease, influencing both its aggressiveness and response to treatment.
Therapeutic interventions
Tumor evolution can naturally occur due to intrinsic genomic instability and stochastic processes. However, therapeutic interventions accelerate and redirect these evolutionary pathways by imposing strong selective pressures. Each treatment modality targets specific vulnerabilities, which inevitably alters the competitive fitness landscape of tumor cells, promoting the emergence of resistant clones and adaptive phenotypes. Understanding these treatment-driven evolutionary dynamics is essential for developing more durable and adaptive therapeutic strategies. Moreover, persister cells, which are transiently drug-tolerant subpopulations within tumors, play a crucial role in cancer progression and therapy resistance. These cells are not genetically resistant but instead exhibit adaptive mutability, allowing them to survive under therapeutic stress and acquire mutations that can lead to permanent resistance. The mechanisms driving this phenomenon include changes in DNA repair pathways, the upregulation of error-prone DNA polymerases and microsatellite instability (MSI).252
Chemotherapy
Chemotherapy remains a cornerstone of cancer treatment and the only systemic treatment for certain cancer subtypes associated with poor clinicopathologic features. By exerting strong cytotoxic effects on cancer cells, chemotherapeutic agents not only eliminate sensitive subclones but also impose intense selective pressures that reshape the clonal landscape of tumors, inadvertently promoting the survival and expansion of resistant populations. In breast cancer patients, following paclitaxel treatment, the E545K mutation of the PIK3CA gene became significantly enriched, with the mutation allele frequency increasing from 14% prior to treatment to 34% posttreatment.253 This mutation promotes cell proliferation and inhibits apoptosis by activating the PI3K/AKT signaling pathway, ultimately leading to resistance to paclitaxel.254 Moreover, whole exome sequencing of tumor biopsies collected before treatment, at the point of therapy switch, and after sequential treatment with epirubicin and docetaxel monotherapy revealed that docetaxel induced a profound redistribution of subclones in breast cancer.255 This redistribution was characterized by the regression of sensitive clones and the expansion of clones carrying mutations associated with treatment resistance, including TP53 and PIK3CA, driving tumor recurrence and evolution.256 Two germline mutations in the TEKT4 gene (c. A541G and c. A547G), associated with reduced disease-free survival and overall survival, were significantly enriched in basal-like breast cancer (BLBC) patients postpaclitaxel-based neoadjuvant chemotherapy (NCT).257 Encoding the microtubule stabilizing protein Tektin4, the ectopic expression of TEKT4 mutations antagonizes paclitaxel-induced microtubule stabilization, contributing to increased paclitaxel resistance.257,258,259
In addition, gene fusion also plays an important role in paclitaxel-driven tumor evolution. ABCB1 encodes multidrug resistance protein (MDR1), an ATP-binding cassette member involved in the cellular efflux of chemotherapeutic drugs.260,261 In high-grade serous ovarian cancer (HGSOC) and breast cancer, a study found that the ABCB1 gene enhances promoter-driven self-expression through transcriptional fusion events with other genes, such as SLC25A40, resulting in overexpression of MDR1.262 MDR1 overexpression reduces the cytotoxicity of drugs by increasing the ability of cells to efflux paclitaxel. After chemotherapy with paclitaxel or other MDR1 substrate drugs, resistant clones harboring the ABCB1 fusion are selectively expanded.262
Notably, paclitaxel-resistant cells are more likely to enter this drug-tolerant persistence state under EGFR-TKI treatment pressure than cells without paclitaxel resistance, eventually evolving into stable resistance.263 With the upregulation of stemness markers such as Nanog and CD133, this process represents a key step in the evolution of tumor cells from paclitaxel resistance to stable resistance to EGFR-TKIs.264,265,266 However, this drug-tolerant persistence state can be reversed after drug removal.267 While paclitaxel-resistant tumor cells were proven to maintain their resistant phenotype by regulating FOXO3a,264 this stable resistance to EGFR-TKIs following paclitaxel resistance does not require known resistance-promoting mechanisms such as MET amplification and KRAS G12 missense mutation. It is primarily driven by a range of cellular phenotypic changes and nongenetic mechanisms, including stemness enrichment, metabolic rewiring, and alterations in apoptosis regulation.
Platinum-based chemotherapeutics, another key component of cancer chemotherapy, introduce a different set of evolutionary dynamics by targeting DNA integrity via the formation of intra- and interstrand crosslinks. A study demonstrated that the newly defined cisplatin-specific single nucleotide variation (SNV) and short insertion/deletion mutations of either base to thymine at AG putative intrastrand crosslinks are causes of the reversion of BRCA2 mutations in emerging cisplatin-resistant clones.268 Cisplatin-specific mutation signatures with significantly increased C>A and T>A mutations were observed in esophageal squamous cell carcinoma (ESCC), urothelial carcinoma (UC), muscle-invasive bladder cancer and high-grade serous ovarian cancer (HGSOC), and the mutation frequency was positively correlated with drug dose in ESCC.86,269,270 These studies suggest that cisplatin-based therapy can directly induce resistance-causing mutations.
Moreover, patients with HGSOC carrying BRCA1/BRCA2 mutations are particularly sensitive to platinum-based drugs due to the lack of homologous recombination repair ability, resulting in a longer recurrence-free survival period.271 However, some patients exhibit platinum resistance after relapse, and restoration of BRCA1 and BRCA2 function was identified as a resistance mechanism.269 Secondary mutations can restore wild-type amino acid sequences, and in-frame deletion of the original mutation can restore at least partial protein function.269,272 Platinum has also been found to induce intracellular DNA damage, upregulate the COL6 (collagen VI) gene, and activate the adhesion signaling pathway associated with the extracellular matrix (ECM) in HGSOC.273 After carboplatin chemotherapy in HGSOC, the PBX1 gene is significantly upregulated in recurrent ovarian cancer, directly promoting the development of tumor stem-like characteristics, including self-renewal, colony formation, and increased resistance to platinum-based chemotherapy.274 PBX1 activates the STAT3 pathway by directly binding to the STAT3 promoter region, and its downstream genes include ATP binding cassette transporters such as ABCA1 and ABCA3, which are related to drug metabolism, reduce cell toxicity and result in resistance by excreting platinum-based drugs. Targeting PBX1 or its downstream signaling pathway JAK2/STAT3 can significantly enhance sensitivity to carboplatin, suggesting a potential new target for delaying carboplatin resistance. In addition, platinum-based chemotherapy can also significantly upregulate the phosphorylation of the proapoptotic protein BAD, thereby inhibiting apoptosis and promoting tumor cell survival.275
Antimetabolites, including nucleoside analogs and pyrimidine analogs, exert selective pressure by targeting nucleotide metabolism. Common nucleoside analogs in tumor treatment include cytarabine and gemcitabine, while common pyrimidine analogs are fluorouracil and capecitabine. In gemcitabine-resistant breast cancer, the cytosine deaminase (CDA) gene is significantly overexpressed, which promotes the inactivation of gemcitabine metabolism and leads to treatment resistance.276 Overexpression of CDA is caused by downregulation of miR-484 and targeting of the 3’ untranslated region (3’-UTR) of CDA after treatment with gemcitabine. In addition, under the pressure of gemcitabine treatment, subclones with high CDA expression also enhance adaptive advantages through antioxidant stress mechanisms and are selectively expanded during the treatment process.276 Through the use of a barcoding system to label and track tumor clones, it was found that gemcitabine treatment changed the clonal composition of pancreatic cancer.277 Approximately 6% of the clone subpopulations were amplified after treatment, while approximately 20% were eliminated. Gemcitabine-resistant clones exhibit lower levels of DNA damage and stronger DNA repair capabilities, including enhanced activity of DNA damage repair pathways such as RAD51 aggregation. Targeted DNA repair pathways, such as the use of the ATR inhibitor AZD-6738, can enhance the efficacy of gemcitabine.277 Transcriptome changes such as oxidative phosphorylation and enhanced activity of MYC/E2F target genes have also been observed in different pancreatic cancer subclones resistant to gemcitabine. Moreover, through the analysis of cell-free DNA (cfDNA) in liquid biopsy, it was found that new gene mutations, including TP53, EZH2, PIM1, MYC and other genes, appeared in patients with diffuse large B-cell lymphoma (DLBCL) treatment resistance during rituximab-cyclophosphamide-doxorubicin-vincristine-prednisone (R-CHOP) chemotherapy.278 The expansion of R-CHOP-resistant DLBCL clones was also manifested in the activation of the PI3K/AKT and JAK-STAT3 signaling pathways after treatment.279
Targeted therapy
Compared to chemotherapy, targeted therapies provide a more precise approach by focusing on specific molecular alterations that drive tumor growth and survival. However, despite their specificity, targeted agents are not exempt from tumor evolution by introducing selective pressures that favor the emergence of resistant clones harboring secondary mutations, bypass signaling, or phenotypic plasticity.
Among various targeted therapies, epidermal growth factor receptor (EGFR) inhibitors represent a key advancement in the treatment of cancers driven by aberrant EGFR signaling, such as non-small cell lung cancer (NSCLC). The evolutionary dynamics induced by EGFR inhibitors have been extensively studied in lung cancer. Gefitinib, an EGFR tyrosine kinase inhibitor (TKI), is widely used to treat EGFR mutation-positive NSCLC patients. The T790M mutation is the primary mechanism of acquired resistance in NSCLC during EGFR inhibitor therapy, occurring in 50–60% of patients.264,280 Research has shown that T790M mutations arise through two pathways: selection of preexisting T790M clones and genetic evolution of drug-tolerant cells. The former involves the expansion of rare T790M mutant clones that exist at low frequencies before treatment, while the latter refers to genetic changes in drug-tolerant cells that acquire the T790M mutation over time. Gefitinib-resistant cells exhibit epigenetic alterations, such as decreased BIM protein levels, which prevent apoptosis and enable the evolution of resistant clones harboring the T790M mutation during prolonged treatment.264
Resistance to EGFR inhibitors in mCRC patients is associated with the emergence of mutations in downstream effectors of the EGFR pathway, such as KRAS, NRAS, and BRAF, which lead to constitutive activation of the MAPK pathway, bypassing the need for EGFR signaling.281 Additionally, mutations in the EGFR extracellular domain (ECD), which impair antibody binding, can also contribute to resistance.282 CtDNA profiling has become a crucial tool for monitoring tumor evolution during EGFRi treatment. Studies have shown that mutations in genes such as KRAS, NRAS, MET, ERBB2, FLT3, and MAP2K1 are commonly detected in ctDNA in patients with acquired resistance to EGFR blockade.281 These alterations facilitate tumor progression by activating compensatory signaling pathways, allowing cancer cells to thrive despite EGFR inhibition. In colorectal cancer (CRC), EGFR/BRAF inhibition downregulates mismatch repair (MMR) and homologous recombination (HR) repair pathways while simultaneously upregulating error-prone polymerases such as DNA polymerase eta and DNA polymerase zeta. This shift toward error-prone DNA repair mechanisms enhances mutability and genomic instability, promoting the survival of persister cells.252,283,284
While the EGFR T790M mutation is the main cause of resistance to first- and second-generation EGFR inhibitors, third-generation EGFR TKIs, such as osimertinib, specifically target this mutation.264,285 Despite high response rates, disease progression inevitably occurs in patients treated with osimertinib. Resistance mechanisms to osimertinib are heterogeneous and can be categorized into on-target mechanisms (e.g., EGFR-mediated resistance, such as the acquisition of C797S) and off-target mechanisms (e.g., bypass signaling via MET amplification or histologic transformation).280,286 Additionally, the EGFR C797S mutation has been shown to induce cross-resistance to all third-generation TKIs by preventing drug binding to the EGFR active site.287 Other resistance mechanisms include mutations in the RAS gene, MAPK gene amplification, and the inactivation of ERK negative regulatory factors, which have garnered widespread attention in third-generation EGFR TKI resistance research.162 Epithelial–mesenchymal transition (EMT) and SCLC transformation, as mechanisms of resistance to first-generation TKIs, may also be related to resistance to third-generation EGFR TKIs. Additionally, a study analyzing 45 patients with resistance to osimertinib revealed that 20% (9 cases) developed new secondary driver mutations.288 These mutations included gene fusions such as KIF5B-RET, STRN-ALK, and FGFR3-TACC3; point mutations such as BRAF V600E and KRAS G12D/G12V/G12R; and other alterations, including PIK3CA E545K and E726K. These findings suggest that tumors can acquire novel driver gene alterations during osimertinib treatment, contributing to resistance and increasing tumor heterogeneity. In EGFR-mutant NSCLC models, treatment with EGFR inhibitors induces activation of the WNT/β-catenin signaling pathway and significant upregulation of the SERPINE1 gene in resistant tumor cells.289 SERPINE1 promotes tumor cell invasion and an immunosuppressive state by regulating the plasminogen activation pathway, which is associated with resistance to EGFR inhibitors and shorter progression-free survival. At advanced stages of EGFR inhibitor treatment, a marked upregulation of genes associated with squamous cell differentiation (e.g., KRT16, KRT14) has been observed, suggesting histologic transformation from adenocarcinoma to squamous cell carcinoma.
Poly (ADP-ribose) polymerase (PARP) inhibitors have emerged as a promising class of targeted therapies, particularly for cancers harboring homologous recombination deficiency (HRD). However, despite their initial efficacy, resistance to PARP inhibitors is a growing concern. Transcriptomic analysis has revealed significant changes in gene expression following olaparib treatment in TNBC, particularly genes related to the DNA damage response (DDR) pathway, such as RAD51 and BRCA1.290 In initial samples from prostate cancer patients treated with olaparib, mutations associated with HRD, such as BRCA1, BRCA2, and PALB2 mutations, were detected.291 During disease progression and the emergence of resistance, secondary mutations in these genes were observed, restoring the open reading frame of the genes and reactivating DNA repair functions, which ultimately led to resistance.290,292 The emergence of new mutations suggests that resistance results from clonal evolution driven by drug selection pressure. Therefore, the significant restoration of HRD-related gene mutation expression induced by olaparib treatment, representing epigenetic reversal, may be related to resistance. The core mechanism of this resistance lies in the tumor’s ability to restore DNA repair capability through gene repair, thereby evading the cytotoxic effects of PARP inhibitors.293 Furthermore, BRCA2 reversion mutations in different metastatic lesions exhibited notable spatial heterogeneity, with some mutations specific to individual metastases, while others were present in multiple metastatic sites. This highlights the strong clonal selection pressure exerted by the drug. Studies suggest that PARP inhibitor treatment not only induces resistance through gene mutations but also may cause phenotypic and functional changes via nongenetic mechanisms. Single-cell RNA sequencing (scRNA-seq) showed that during olaparib treatment, tumor cells undergo a dynamic transition from an epithelial to a mesenchymal-hybrid state (EM-hybrid), which further develops into a stable mesenchymal phenotype posttreatment.290 The transcription factors SNAI2 and TWIST2 played a sustained role in this transition.
CDK4/6 inhibitors, such as palbociclib, ribociclib, and abemaciclib, have emerged as effective treatments for various cancers, particularly hormone receptor (HR)-positive breast cancer. Following treatment with palbociclib and fulvestrant, the RB1 gene exhibited inactivating alterations, including stop codon and frameshift mutations, which facilitated the mutation of drug-resistant clones.294 No RB1 mutation was observed with fulvestrant monotherapy. New PIK3CA mutations, mainly hotspot mutations such as E542K, were also detected in drug-resistant patients after treatment.294 Studies have shown that palbociclib combined with fulvestrant treatment amplified subclones carrying RB1, PIK3CA and ESR1 mutations through selective pressure.295 This suggests that PIK3CA and ESR1 mutations may cooperate with the CDK4/6 signaling pathway to drive the expansion of drug-resistant clones. In HER2-positive breast cancer, combination treatment with a CDK4/6 inhibitor and an anti-HER2 antibody rapidly causes drug resistance in tumors.296 Single-cell RNA sequencing (scRNA-Seq) and tumor microenvironment (TME) analysis indicated immunosuppressive alterations characterized by substantial infiltration of immature myeloid cells (IMCs) into drug-resistant tumors, elevated levels of immunosuppressive cytokines such as CXCL1 and TGFβ3, and reduced infiltration of T cells and NK cells within the TME.
HER2-targeted monoclonal antibodies include trastuzumab and pertuzumab. Through comparative analysis of the genomes of HER2-positive advanced gastric cancer patients before and after trastuzumab treatment, it was found that ERBB4 and FAT4 mutations were the most common treatment-acquired mutations, significantly reducing progression-free survival (PFS).297 Further research suggests that ERBB4 mutations may promote drug resistance by activating alternative signaling pathways, such as bypassing HER2, while FAT4 mutations may promote cancer cell proliferation and invasion by inactivating YAP1 inhibition.297
Vascular endothelial growth factor (VEGF) inhibitors, such as bevacizumab, function by blocking angiogenesis and limiting tumor blood supply. After receiving treatment with the antiangiogenic drug bevacizumab, POLR1D amplification at the 13q12.2 site was detected in resistant colorectal cancer patients, leading to upregulation of VEGFA and EREG expression.298 VEGFA is an important member of the vascular endothelial growth factor family and is directly involved in resistance to anti-VEGF drugs. As the disease progresses, clones with amplification of the 13q12.2 site become the dominant clone and are directly associated with treatment resistance. In addition, CREBBP and FBXW7 mutations were detected in patients with metastatic colorectal cancer during the resistance phase after receiving anti-VEGF chemotherapy.299 Both of these genes are tumor suppressor genes, and their mutations lead to the survival and expansion of cells under low oxygen conditions after anti-VEGF therapy, thereby driving the development of drug-resistant clones.
Endocrine therapy
While endocrine therapies have revolutionized the management of hormone-dependent tumors, their clinical success is frequently undermined by the emergence of resistance. Different classes of endocrine therapies, including anti-estrogen agents, anti-androgen agents, and aromatase inhibitors, exert selective pressures on tumor cells by disrupting hormonal signaling pathways essential for tumor growth.
Research on anti-estrogen therapies primarily focuses on hormone receptor-positive/HER2-negative (HR+/HER2-) advanced breast cancer. Tumor genome evolution and treatment resistance are common during combination therapy with antiestrogenic drugs (such as bazedoxifene or fulvestrant) and the CDK4/6 inhibitor palbociclib. Mutations in the ESR1 gene, particularly the Y537S mutation, and PIK3CA mutations are well-established drivers of fulvestrant resistance and play a dominant role in posttreatment resistance mechanisms.294,295,300 These mutations significantly impact progression-free survival (PFS). To address fulvestrant resistance, researchers have proposed novel personalized therapeutic strategies, such as combining PI3K inhibitors targeting PIK3CA mutations or next-generation selective estrogen receptor degraders (SERDs) targeting ESR1 mutations with palbociclib and fulvestrant.294,295 These combinations may improve outcomes and optimize treatment efficacy. Selective amplification of the PIK3CA activating mutation and ESR1 Y537S mutation was also detected in some patients during treatment, which was proven to be directly related to fulvestrant resistance.294,295 Moreover, a separate study identified APOBEC3B (A3B) as a key driver of genomic mutations and resistance development in ER+ breast cancer under tamoxifen treatment.301 Through its C-to-U DNA deamination activity, A3B actively promotes tumor evolution and contributes to therapeutic resistance.
Aromatase inhibitor (AI) therapy drives the enrichment and expansion of KRAS and ESR1 (estrogen receptor α) mutations through selective pressure, significantly contributing to resistance in HR+ advanced breast cancer.302,303 Among patients with disease progression, 15.4% harbor RAS mutations, including KRAS, HRAS, and NRAS, with multiclonal KRAS mutations, such as G12V and G12C, being particularly prominent.302 ESR1 mutations predominantly involve common ligand-binding domain alterations, such as Y537S and D538G, which result in ligand-independent activation of the estrogen receptor (ER). This enables tumor cells to survive and proliferate in estrogen-deprived conditions, leading to AI therapy failure.303 Additionally, in ER+ breast cancer, resistance to combined treatment with the CDK4/6 inhibitor ribociclib and the endocrine therapy letrozole is associated with a marked reduction in ESR1 expression, including loss of heterozygosity (LOH) of ESR1 in some patients.304 Resistant tumors also exhibit a shift in signaling pathways, transitioning from ERK signaling in sensitive tumors to JNK signaling.304 Upregulation of JNK signaling is correlated with the proliferation of resistant tumor cells, while ERK signaling remains highly active in treatment-sensitive tumors. Additionally, CYP19A1 amplification has been identified as a resistance-associated genomic alteration specifically selected under AI therapy.305 CYP19A1 amplification significantly enhances aromatase activity, enabling tumor cells to autonomously produce estrogen through endocrine pathways. This, in turn, activates ERα-dependent gene expression, allowing tumor cells to evade immune surveillance and therapeutic effects. A multitargeted therapeutic strategy addressing the ER, PI3K, and RAS pathways may overcome treatment resistance and improve patient outcomes.302
Anti-AR therapies are widely used to treat castration-resistant prostate cancer (CRPC). However, AR-targeted treatments such as abiraterone and enzalutamide have been shown to selectively induce and sustain the expression of AR splice variant 7 (AR-V7), which contributes to therapy-associated resistance.306 Notably, taxane-based chemotherapy may partially alleviate this resistance by reducing the overall number of circulating tumor cells (CTCs) or by disrupting the AR signaling axis, thereby decreasing AR-V7 expression.306 The evolution of the AR gene and associated pathways also plays a pivotal role in the development of resistance to AR-targeted therapies.307 AR signaling inhibitors exert strong selective pressure, driving significant genomic alterations in CRPC patients, including AR gene amplification, mutations, and rearrangements.308 During treatment, the AR gene copy number increased by an average of 50%. Two ligand-binding domain (LBD) mutations, L702H and F877L, have been identified, with the former linked to resistance to abiraterone and the latter to enzalutamide. Furthermore, LBD truncating rearrangements in the AR gene have been observed in resistant patients, allowing ligand-independent activation of AR and further promoting therapeutic resistance. Acquired resistance to anti-androgen therapies is also associated with additional mutations or deletions in TP53, PTEN, and RB1, suggesting their involvement in tumor evolution and resistance mechanisms.308
Immunotherapy
Immunotherapy, particularly immune checkpoint inhibitors, adds another dimension to this dynamic by targeting immune evasion rather than tumor-intrinsic pathways. Immune checkpoint inhibitors (ICIs), such as anti-PD-1/PD-L1 and anti-CTLA-4 therapies, have ushered in a new era of cancer treatment by enhancing antitumor immunity. When treated with the immune checkpoint inhibitors nivolumab and pembrolizumab, melanoma, NSCLC and glioblastoma (GBM) patients who respond to treatment exhibit selective killing of tumor cells with high neoantigen loads, leading to the enrichment of low-immunogenic subclones.309,310,311,312,313 This process is referred to as clonal contraction or immunoediting. In contrast, tumors from resistant patients show mutational expansion, indicating that therapeutic pressure has triggered an increase in tumor clonal diversity. Similarly, in immune checkpoint inhibitor nonresponders, accelerated mutation accumulation posttreatment is observed, manifested as linear evolution.314 This suggests that treatment pressure may contribute to resistance by facilitating the expansion of subclonal populations.
Moreover, in nivolumab-resistant melanoma patients, the loss of CDKN2A and an increase in B2M mutations are significantly observed.309 Frameshift mutations in B2M, along with heterozygous loss, are a typical mechanism of acquired resistance to PD-1 therapy.309,310,311 In NSCLC, reductions in the copy number of B2M and antigen-presentation genes (such as TAP1, TAP2, and HLA-A) have been shown to be associated with nivolumab resistance.311 Additionally, in some nivolumab-resistant clear cell renal carcinoma (ccRCC) patients, HLA loss of heterozygosity (HLA LOH) and deletions of CDKN2A/B copy numbers have been observed, significantly reducing antigen presentation ability and promoting tumor immune evasion.312 In regions with a significant immune response posttreatment, HLA LOH or CDKN2A/B loss is typically absent. PTEN mutations are significantly enriched in GBM nonresponders to anti-PD-1 therapy.314 Mutated PTEN activates the PI3K-AKT signaling pathway, promoting immune suppression and reducing T-cell infiltration. Within the immune microenvironment, PTEN mutations are also associated with the accumulation of macrophages and neutrophils, potentially enhancing immune evasion capabilities.314 After anti-PD-1 treatment, upregulation of PD-L1 expression and synergistic increases in other checkpoint molecules, such as LAG-3 and TIM-3, observed in drug-resistant samples further weakened the immune response.313 In nonresponders, there is an accumulation of immunosuppressive cells such as Tregs and myeloid-derived suppressor cells (MDSCs) within the TME posttreatment.315 While T-cell clonal diversity increases significantly, there is a lack of effective antitumor immune activity, suggesting that adaptive changes in the immune microenvironment may limit the effectiveness of immunotherapy.
Hyperprogressive disease (HPD) represents an unexpected, rapid tumor progression following immune checkpoint inhibitor (ICI) therapy.316 Despite the primary goal of ICIs being to boost antitumor immunity, a subset of patients experiences HPD, where tumors grow more aggressively after treatment. The mechanisms behind HPD remain complex and multifactorial. One of the key mechanisms contributing to HPD is the alteration of the tumor microenvironment (TME). ICIs can shift the immune landscape, leading to the expansion of immunosuppressive cells, such as myeloid-derived suppressor cells (MDSCs), regulatory T cells (Tregs), and macrophages, which may promote tumor progression rather than inhibit it.316,317 These cells can suppress T-cell activity, facilitating tumor growth. Moreover, T-cell exhaustion or dysfunctional T-cell phenotypes have been linked to HPD. The Fc domain of various therapeutic antitumor immunoglobulin Gs (IgGs) plays a pivotal role in their therapeutic efficacy through its interaction with Fcγ receptors (FcγR), which triggers tumor cytotoxicity.318 PD-1 monoclonal antibodies (mAbs) engage T cells via the Fab domain, while also recruiting natural killer (NK) cells and macrophages through the Fc domain. Consequently, NK cells and macrophages can eliminate T cells via antibody-dependent cell-mediated cytotoxicity (ADCC) or antibody-dependent cellular phagocytosis (ADCP), potentially leading to T-cell depletion and facilitating rapid tumor progression.319 Genomic alterations, such as mutations in driver genes and alterations in oncogenic pathways, contribute to HPD. Specific mutations, such as copy number alterations in MDM2/MDM4 and EGFR, have been shown to drive aggressive tumor behavior in response to ICI therapy. Transcriptional upregulation of tumor-related signaling pathways, including IGF-1, ERK/MAPK, and PI3K/AKT, has also been observed.320 Furthermore, disruption in IFN-γ signaling has been implicated in HPD.317 While IFN-γ is typically involved in promoting antitumor immunity, its upregulation in certain contexts during ICI therapy can paradoxically enhance tumor evasion mechanisms. This occurs through the induction of PD-L1 expression on tumor cells, leading to resistance to immune surveillance.321
Radiotherapy
Radiotherapy is a cornerstone of cancer treatment, influencing tumor evolution by imposing selective pressures that shape tumor growth and response dynamics. The effectiveness of radiotherapy often leads to the expansion of resistant subclonal populations, contributing to relapse and therapeutic resistance. Recent studies have revealed several key signaling pathways and molecular mechanisms that mediate tumor evolution in response to radiotherapy.
A critical aspect of tumor evolution following radiotherapy involves alterations to the tumor microenvironment, particularly immune cell dynamics. For instance, radiation can enhance the accumulation of myeloid-derived suppressor cells (MDSCs), which play a pivotal role in immune suppression and tumor progression.322 By interacting with T cells and natural killer (NK) cells, these cells exert immunosuppressive effects that contribute to tumor survival and metastasis. Moreover, radiotherapy induces genetic mutations and genomic instability, fueling the clonal evolution of tumors. In undifferentiated pleomorphic sarcoma (UPS), radiotherapy has been shown to drive a shift toward subclonal selection, where specific mutations, such as those in calcium signaling pathways, are linked to radiosensitivity.323 Radiation causes significant changes in subclone abundance, with subclone contraction correlating with alterations in calcium signaling pathways. Targeting calcium transporters has been shown to enhance UPS cell radiosensitivity, highlighting the crucial role of calcium signaling in radiotherapy resistance.323 Therefore, radiotherapy not only induces DNA damage and mutations in tumor cells but also plays a significant role in tumor evolution through immune modulation and selective subclone expansion. A deeper understanding of the signaling pathways involved in radiotherapy-induced tumor evolution, including immune regulation and genomic instability, is essential for developing targeted strategies to enhance treatment efficacy and minimize relapse. Moreover, in lung tumors during radiotherapy, a patient-specific model based on a partial differential equation was built to predict the evolutionary trajectory, showing a satisfying agreement between the predicted tumor contours and those drawn by an expert.324
In conclusion, treatment-induced tumor evolution plays a central role in the development of therapeutic resistance, complicating the efficacy of various cancer treatments. Chemotherapy, immunotherapy, endocrine therapy and targeted therapies all exert selective pressures that drive tumor cells to adapt through genetic alterations, clonal expansion, and the activation of alternative signaling pathways. In chemotherapy, resistance often arises through mutations in drug target genes or by upregulating drug efflux pumps. Immunotherapy resistance is primarily driven by immune escape mechanisms, such as tumor-induced immune suppression. In targeted therapies, mutations in key oncogenes such as EGFR or PI3K lead to the emergence of resistant clones, while endocrine therapies for HR+ breast cancer are hampered by mutations such as ESR1. These evolving tumor dynamics underscore the need for personalized and combination treatment strategies that account for tumor heterogeneity and adaptive resistance, ultimately optimizing therapeutic efficacy and improving patient outcomes.
Environmental factors
Environmental factors are significant external regulators of tumor evolution, influencing cancer progression, metastasis, and therapy resistance. These factors include a variety of physical, chemical, and biological elements, such as diet, air pollution, UV radiation, and infection, each of which can induce specific genetic and epigenetic changes in tumor cells. Environmental exposure creates selective pressures that shape tumor evolution, driving the selection of clones that are better suited to survive and thrive in the given microenvironment. This section will focus on the impact of environmental factors on tumor evolution, highlighting key molecular pathways and mechanisms through which these factors influence cancer progression.
UV radiation is one of the most well-known environmental carcinogens and is primarily associated with skin cancers, such as melanoma.325 UV radiation induces DNA damage in skin cells, leading to the formation of pyrimidine dimers and the activation of DNA repair mechanisms. However, the accumulation of unrepaired DNA damage can promote genetic mutations and drive tumor evolution. UV radiation-induced DNA damage is primarily repaired through nucleotide excision repair (NER) and other repair pathways. However, defects in these repair mechanisms, particularly in genes such as XPC and ERCC2, increase the likelihood of mutations, which may contribute to the development of skin cancers.326 UV radiation also activates cellular stress responses, including the activation of p53, a key tumor suppressor protein involved in the DNA damage response and cell cycle regulation. However, chronic exposure to UV radiation can lead to mutations in the p53 gene itself, which may allow for the accumulation of additional mutations and the progression of cancer. Furthermore, UV-induced DNA damage is often associated with the activation of oncogenic pathways such as the MAPK and PI3K/AKT signaling pathways, which promote cell survival and proliferation. These pathways can be upregulated in response to DNA damage, creating a feedback loop that supports tumorigenesis and drives the evolution of resistant subclones. UV radiation also influences the immune system by suppressing local immune responses, thereby facilitating tumor survival and growth. This immune suppression is partly mediated by the induction of regulatory T cells (Tregs), which inhibit antitumor immune responses and support the evolution of immune-resistant tumor cells. Thus, UV radiation is a powerful environmental factor that induces DNA damage and promotes tumor evolution by selecting for mutations that enhance tumor survival and resistance to immune surveillance. It has also been found to shape dendritic cell leukemia transformation in the skin by promoting plasma-cytoid dendritic cell (pDC) expansion and mutagenesis.327 The impact of UV radiation on tumor progression underscores the importance of protecting against environmental factors that promote genetic instability and drive cancer development.
Chemical carcinogens, such as those found in tobacco smoke, industrial chemicals, and pollutants, are another major environmental factor influencing tumor evolution. These substances induce DNA damage by forming adducts with DNA bases, leading to mutations that can contribute to tumorigenesis. One of the most well-studied examples of chemical carcinogenesis is the role of tobacco smoke in the development of lung cancer. Tobacco smoke contains numerous mutagenic compounds, including polycyclic aromatic hydrocarbons (PAHs), which cause DNA damage and mutations that promote tumor evolution.328 Adler et al. investigated how exposure to tobacco carcinogens induces mutations in key cancer-related genes, such as TP53, EGFR, and KRAS, leading to the development of drug-resistant subclones. These mutations are often selected for in response to the selective pressures imposed by the carcinogen, driving the evolution of a genetically diverse tumor population. Additionally, the accumulation of mutations in DNA repair genes, such as those in the base excision repair pathway, further accelerates the rate of mutation and contributes to the generation of subclonal diversity.
Chemical carcinogens also modulate signaling pathways that drive tumor progression. For example, PAHs and other carcinogens can activate the aryl hydrocarbon receptor (AhR), which regulates the expression of detoxification enzymes and can influence tumor growth and metastasis.329 The activation of AhR and other signaling pathways by chemical carcinogens promotes tumor evolution by creating a proinflammatory and prosurvival microenvironment, further facilitating the development of therapy-resistant tumor clones.
Additionally, infections, particularly viral infections, also contribute significantly to tumor initiation and progression.330 For instance, human papillomavirus (HPV) is linked to cervical cancer, while chronic hepatitis B and C infections lead to hepatocellular carcinoma. These infections often cause genomic instability, inflammation, and changes in epithelial cell differentiation, which contribute to the tumorigenic process. Vaccination against HPV has been shown to reduce the incidence of early neoplasias and cervical cancers by inducing regression of precancerous lesions, demonstrating how infection-related carcinogenesis can be prevented at early stages.
Lifestyle habits of patients
Lifestyle habits of individual patients also influence the development of cancer not only through direct genetic alterations but also by promoting a favorable environment for mutated cells to thrive and evolve into invasive tumors.
Smoking and alcohol consumption are two of the most significant lifestyle-related risk factors for cancer.330,331 Smoking introduces mutagenic compounds into the body that directly damage DNA, leading to oncogenic driver mutations. In addition to these mutagenic effects, smoking also acts as a promoter by creating a proinflammatory environment, enhancing cell proliferation, and suppressing immune responses, which collectively promote tumor progression. For example, smoking is linked to several cancers, including lung, oral, and esophageal cancers, where inflammatory markers such as cytokine production in lung myeloid cells play a pivotal role. Alcohol, another major lifestyle factor, affects cancer risk through hormonal alterations (such as increased estrogen levels) and through changes in metabolic processes.332 Alcohol consumption has been associated with a decrease in immune function, particularly through the reduction of CD8 + T-cell responses, thus facilitating tumor growth.
Dietary factors are emerging as important contributors to tumor evolution. Certain diets, particularly those high in fat, sugar, and processed foods, can create a metabolic environment that supports tumor growth and progression. These dietary factors can influence signaling pathways involved in metabolism, inflammation, and DNA repair, driving tumor evolution. For example, a high-fat diet has been shown to activate the mTOR signaling pathway, which regulates cell growth, metabolism, and survival. This activation promotes tumor growth by enhancing nutrient uptake and increasing cell proliferation.333 A high-fat diet also promotes cancer evolution by inducing gut microbiota-mediated leucine production and myeloid cell differentiation.334 Similarly, diets rich in antioxidants have been shown to suppress oxidative stress, thereby reducing DNA damage and slowing the rate of mutagenesis. Conversely, diets that promote inflammation, such as those high in processed foods and red meat, can exacerbate tumor progression by activating proinflammatory cytokines and signaling pathways such as NF-κB, further driving the evolution of therapy-resistant tumor cells.333 Thus, dietary factors can influence tumor evolution by modulating key signaling pathways, such as those involved in metabolism, inflammation, and DNA repair, contributing to the development of more aggressive and resistant tumor clones.
