Exploring the Complexity of Pan-Cancer: Gene Convergences and in silico Analyses

Introduction

Cancer may be defined as a set of multifaceted diseases that converge through deregulation of both replicative processes, leading to abnormal cell proliferation, and of molecular mechanisms linked to cell death.

Tumorigenesis primarily occurs due to factors such as mutations in key genes (oncogenes and/or tumor suppressor genes), epithelial–mesenchymal transition process (EMT), emergence of a tumor microenvironment favoring tumor progression, and activation of immune evasion mechanisms, angiogenesis, and metastasis.

In 2020, the global incidence of cancer exceeded 19 million cases, with breast, lung, and colorectal cancers having the highest absolute number of cases, regardless of sex or age, respectively.¹ In the same year, the global mortality rate, regardless of sex or age, reached 51.61%, with lung, colorectal, and liver tumors representing 35.7% of this total.¹

Tumor Classification

Tumors can be categorized as in situ or invasive, corresponding to the so-called benign and malignant cancers, respectively. Tumor classification can also utilize anatomical characteristics, according to the Classification of Malignant Tumors (TNM), which uses tumor extent based on the extent of the primary tumor (T, T1 – T4); the absence or presence of metastasis in adjacent lymph nodes (N, N0 – N2), and the absence or presence of distant metastasis from the primary site (M, M0 – M1). The TNM varies depending on the type of tumor, as exemplified in Table 1 below:

Table 1 TNM Classification of Different Tumor Types

Biomarkers

Although histopathological analyses are important in oncological research, biomarkers have increasingly become the main focus in tumor diagnosis and clinical prognosis. Biomarkers are generally associated with cellular processes broadly related to tumorigenesis and tumor progression, such as uncontrolled proliferation, differentiation and cell migration, in addition to being involved in the deregulation of molecular functions associated with cell adhesion, cell communication and invasiveness² (Table 2).

Table 2 Main Tumor Biomarkers Used in Clinic

Tumor Microenvironment

Recently, the tumor microenvironment has become a major player in tumor studies, regulating neoplastic progression processes,³ with direct action in tumorigenesis and cell migration exerted by fibroblasts, endothelial cells, and adipocytes. These components also act through molecular factors such as Vascular Endothelial Growth Factor (VEGF), in tumor neovascularization and metabolism.⁴ Non-cellular components also directly contribute to the progression and increase in tumor malignancy, with the extracellular matrix (ECM) being the primary component. Composed of fibronectin, collagen, elastin, and laminin, the ECM provides a physical network that forms the main structure of solid tumors.⁵ The process of desmoplasia, in which connective tissue, richly composed of ECM, grows abnormally adjacent to a preexisting tumor, is directly associated with a poor prognosis for patients.⁶

In summary, tumors, regardless of their classification, exhibit a high degree of heterogeneity at their core, relating to the types of genes expressed, associated biomarkers, existing regulatory pathways, and the tumor microenvironment, which varies in components and degrees of differentiation, depending on the tumor in question. However, despite their heterogeneity, tumors converge at various points, from their origin to the stages of neoplastic progression. A new area of study has emerged in tumor research, utilizing contemporary biocomputational tools to correlate hundreds of tumor subtypes and their convergent processes.

Pan-Cancer

Pan-cancer is a new concept aimed at studying the heterogeneity of tumors from a macro perspective, utilizing next-generation sequencing technologies capable of generating extensive databases. By integrating biocomputational technologies, these databases comparatively examine the similarities and differences among various types of cancers, thus facilitating the integration of migration, proliferation, and tumor invasion within a general genetic regulatory network framework.⁷

Biobanks have led to a substantial increase in the number of potential genes that may be related to tumorigenic processes, especially from similar genetic signatures across different tumor types.⁸ Such signatures may be associated with gene substitutions, insertions, or deletions, as shown in Table 3.

Table 3 Genetic Mutations Related to Tumors

In addition to the aforementioned genetic signatures, international consortia propose using many omics data as a method to search for patterns of cellular origin (cell-of-origin) based on analyses of molecular similarities, which relate the origins of tumor cells with the aspects of in situ and invasive tumors; oncogenic processes, which analyze how genetic variants and somatic mutations influence tumor progression; and, also, standard mechanisms associated with key molecular pathways correlated with genes previously established to act in tumor malignancy (MYC, RAS) and cellular processes directed toward ubiquitination, splicing, and energy metabolism.⁹

The goal is to analyze the state of the art concerning tumors, based on the refinement and analysis of public data linked to multi-omics assays in pan-cancer, seeking to identify convergent and divergent patterns concerning gene mutations and differential expression of protein-coding genes, aiming to establish potential new proteins that may be associated with the processes of tumorigenesis and tumor progression.

Methods

A literature search was conducted on PubMed and Google Scholar databases for the last 10 years. We selected 101 articles/public data repositories correlated to basic and clinical research on numerous tumor-related topics: history, epidemiology, intratumoral environment, tumor microenvironment, tumor evasion, tumorigenesis, tumor progression, tumor therapeutics, tumor biomarkers, pan-cancer, tumor cell-of-origin, and extracellular vesicles in the tumor context.

For multi-omics analyses, the following public repositories were used: Pan-cancer Analysis of Whole Genomes of International Cancer Genome Consortium/The Cancer Genome Atlas Program (ICGC/TCGA), China Pan-cancer, and Metastatic Solid Cancers. These repositories provided epidemiological, genetic, and protein prediction information for 20 tumor types. The analyses were conducted using the cBioPortal omics refinement and analysis suite, along with EnrichR for Gene Set Enrichment Analysis and StringDB for protein interaction searches.

Tumor Onset

The current state-of-The-art in tumorigenesis includes the following elements and genes, which are common to all tumors (Figure 1):

• Sequential mutations, which can arise from direct genomic alterations or epigenetic modifications, contributing to a disrupted homeostatic environment and uncontrolled cell proliferation;¹⁰
• Loss of communication with neighboring cells, enabling evasion of the immune system;¹¹
• Ability to express cell survival-promoting factors (or cell immortality through telomerase activation in tumors) and inhibit apoptotic pathways (eg, caspase cascade);¹⁰
• Disruption of normal metabolic pathways that typically involve diverting carbon sources from healthy tissues, increased aerobic glycolysis with elevated lactate production, and a super-stimulus for VEGF-driven angiogenesis;^12,13
• Promotion of inflammatory processes through immune cell recruitment or evasion;¹⁴
• Establishment of invasive processes leading to lymph node and distant metastases.

Figure 1 Genes involved in tumor processes. Main processes related to tumorigenesis and tumor progression and their main mutated genes.

Based on these aspects, cancer can be defined as a complex system of diseases caused by various factors, from deregulation of the mitotic cycle, considering all its inherent complexity (eg, deregulation of proteins from interphase or anaphase checkpoints),^15,16 accumulation of mutations in different genomic regions that cause unavoidable instabilities in the affected cells, to unique mutations in key genes linked to the processes of proliferation, adhesion, and cell migration, such as those encoding p53, CDK1, KIP, BRCA, among others.¹⁷

Another hypothesis proposes that tumor stem cells (TSCs) drive tumorigenesis. Due to their similar characteristics, when compared to normal stem cells, TSCs sustain tumor growth and contribute to resistance against conventional therapies, such as radiotherapy and chemotherapy.¹⁸

Nevertheless, despite their shared origins, tumors exhibit significant heterogeneity, influenced by genetic mutations, the tumor microenvironment, epigenetic factors, and complex molecular mechanisms. Consequently, diagnosing, prognosing, treating, and understanding the global epidemiology of cancer remains a formidable challenge.

Data from the TCGA Project, considering the ratio of synonymous to non-synonymous mutations in human populations, suggest that early-stage cancer on a microevolutionary scale is a neutral process for species fitness. This is indicated by a ratio close to 1 for various tumor-associated genes, contrary to the belief that this ratio would be significantly higher for tumors. However, key genes present extremely high values of this ratio, further characterizing them as a focus of studies in the field of pan-cancer.¹⁹

Thus, an environment of high genetic instability is created, which, coupled with other conditions (altered metabolism, support from tumor-associated cells present in the tumor microenvironment), facilitates tumorigenesis and tumor progression, guided by a microevolutionary sense of selective adaptation of normal cells.

Tumor Morphological Changes

Cell differentiation is essential for creating specialized cells with distinct functions, yet it is not the sole factor driving the complexity of organisms. EMT is a process of transforming polarized epithelial cells into mesenchymal cells. It is essential for neural crest gastrulation and delamination, as well as the generation of motile cells capable of degrading the tissues basal membranes.²⁰ However, due to this characteristic, EMT is strongly associated with tumorigenesis and tumor progression. During EMT, both normal and cancer cells undergo profound physiological changes, driven by transcription factors such as TWIST1/2, Snail, Slug, and ZEB1/2, regulating the expression of cell junction proteins, such as cadherins.²¹

E-cadherin and N-cadherin are responsible for epithelial cell adhesion, signaling, adherent junction formation, and migration. These proteins play essential roles in maintaining cell structure and communication in different tissue types. However, in tumorigenic environments, these transcription factors and proteins facilitate EMT by suppressing E-cadherin expression and upregulating N-cadherin. E-cadherin suppression is primarily caused by the silencing of the CDH1 promoter in E-box regions by Snail, Slug, and Twist,^22,23 which allows cells to lose their adherence to one another and acquire a more mesenchymal property. N-cadherin activation is influenced by factors linked to E-cadherin, such as Twist and the methyltransferase KMT5A. These proteins interact to induce N-cadherin promoter region through H4K20 methylation,²⁴ facilitating the cell transition to a mesenchymal phenotype, thereby promoting migration and invasion.²⁵

Classical Tumorigenesis-Associated Molecular Pathways

Tumorigenesis-associated molecular pathways are essential for understanding tumor initiation and progression. These intricate cell communication networks regulate normal cell growth, proliferation, survival, and differentiation, but their subversion drives cancer onset and progression.²⁶

Extracellular-Regulated Kinase/Mitogen-Activated Protein Kinase Pathway

The extracellular-regulated kinase/mitogen-activated protein kinase (ERK/MAPK) pathway is central to regulating cell proliferation, differentiation, apoptosis, and cell stress responses.²⁷ Its activation occurs mainly through binding of s and peptide growth/differentiation factors to plasma membrane tyrosine kinase receptors (TKRs). This engagement triggers the signaling pathway by activating small GTPases, such as the Ras protein, which then uses a second messenger (G protein) to activate Raf proteins. These, in turn, initiate a phosphorylation cascade, first of MEK1/2, at their serine residues, and then of ERK1/2 through direct action of MEK1/2, facilitating translocation of ERK1/2 to the nucleus.²⁸

ERK1/2 translocation to the nucleus carries out various effector actions linked to cell proliferation, migration, and differentiation, such as phosphorylation of transcription factors c-Jun, c-Fos, and c-Myc, regulation of ribosomal kinases, negative feedback of factors linked to the pathway itself, as well as induction of gene coding for proteins associated with EMT, the cytoskeleton, and the cell cycle.²⁹

Phosphatidylinositol 3-Kinase/Protein Kinase B Pathway

The phosphatidylinositol-3-kinase (PI3K/Akt) signaling pathway is a vital player in intracellular signaling, influencing not only cell proliferation but also the regulation of cell survival, stress response, growth, and metabolism, as well as acting as an effector in cytoskeletal reorganization under various conditions.³⁰ Numerous oncogenes and peptide growth factors are associated with its activation, which is also directly modulated by epigenomic modifications.³¹ In tumors, its activation occurs through two primary mechanisms linked to tyrosine kinase receptor activation, while its overactivation is associated with PTEN tumor suppressor gene inhibition.^32,33 Several stages are involved in PI3K/Akt pathway activation. Peptide growth factor stimuli activate PIK protein subunits through PIP3 synthesis by phosphorylating PIP₂. Consequently, protein kinase B (PKB)/Akt, through its active site, binds to PIP₃, enabling PDK1 phosphorylation. This modification is crucial for activating mTORC1, promoting gene expression processes and cell proliferation.³⁴

Wnt/β-Catenin Pathway

The Wnt pathway has two branches, known as canonical and non-canonical. The non-canonical one is independent of factors linked to β-catenin, while the canonical one involves β-catenin translocation to the nucleus to activate transcription factors.³⁵ The Wnt pathway controls cell proliferation, and the non-canonical branch uniquely regulates cell polarity and tumorigenic migration.³⁶ Wnt ligands in the extracellular environment activate the canonical branch. In their absence, an off-state prevails, directing β-catenin to proteasomal complexes and preventing cytoplasmic accumulation. In an on-state, Wnt ligands bind to Frizzled receptors and other ligands (such as proteins related to low-density lipoproteins 5 and 6 - LRP5/6), phosphorylating β-catenin proteasome complex groups and thus preventing its degradation.³⁷ In the non-canonical pathway, Wnt ligands connect to Frizzled receptors along with co-receptors of tyrosine kinase proteins, initiating a reaction cascade involving GTPases, Ras-like proteins (RhoA), and associated kinase pathways (ERK/JNK/MAPK),³⁸ with effector actions on cytoskeletal rearrangement, cell motility, and polarity.

Tumor Biomarkers

Biomarkers are molecules frequently associated with tumorigenesis and tumor progression, playing a crucial role in regulating malignancy through various molecular mechanisms. They encompass a wide range of compounds, including proteins, glycans, glycoproteins, genetic variants, enzymes, and long non-coding RNAs linked to cancer progression³⁹ (Table 4).

Table 4 Main Types and Specific Biomarkers in Tumors

At the forefront of biomarker research are long non-coding RNAs, such as MALAT1 and HOTAIR, which have been implicated in various cancer types due to their ability to regulate gene expression and influence processes like cellular proliferation and migration. Additionally, the enzyme butyrylcholinesterase (BuChE) is notable in this context. Low levels of BuChE have been associated with poorer prognoses in colorectal cancer cases, suggesting that the dysregulation of its activity may adversely affect disease progression.⁴⁰

While some biomarkers exhibit similarities across different tumor types, many are tissue-specific, making the identification of these biomarkers critical for clinical research. The association of biomarkers with unfavorable prognoses stimulates the development of targeted therapies aimed at improving patient survival and clinical outcomes.

Angiogenesis and Tumor Metastasis

Angiogenesis is a process in which tumor vascular networks emerge and integrate with the body’s vascular system, providing macro and micronutrients necessary for tumor progression. This process begins with the association of endothelial cells with the tumor, followed by the development of a complex capillary network, which integrates with the normal capillary vessels, being essential for tumor growth and subsequent metastasis.⁴¹ Typically, quiescent endothelial cells are activated by elevated levels of pro-angiogenic factors such as VEGF. Following activation by VEGF, other critical components are produced, including factors related to platelet emergence and VEGF receptors (VEGFR). The process proceeds with Notch pathway inhibition in the vascular system, stabilizing and modulating natural angiogenesis. The destabilization and loss of control of the angiogenesis process reshape the vascular system in the tumor region, giving rise to new vessels associated with the tumor.⁴²

Angiogenesis is the primary mechanism in the metastatic process, being associated with tumors classified as malignant. Metastasis is defined as a process in which a primary tumor, following immediate integration with local lymph nodes, spreads to extra local spaces through cell evasion and inhibition of antitumorigenic factors.⁴³

Metastasis initiates when single cells or cell clusters invade nearby lymph nodes or disseminate to distant tissues and organs via tumor-associated vessels.⁴⁴ Tumor cells rapidly adapt to new environments due to their genomic instability. This adaptation involves physical adaptations, such as cytoskeletal and plasma membrane remodeling, and gene expression of proteins involved in desmosomal processes and overall cell adhesion.⁴⁵ Hypoxia- and anoxia-resistant genes are also involved, such as Hypoxia-Inducible Factor (HIF)-induced signaling cascade and Hypoxia-Response Element (HRE) genes, which stimulate metabolism based on anaerobic glycolysis in tumor cells.⁴⁶ Recent investigations have also highlighted the role of tumor-associated macrophages that promote angiogenesis through lactate-secreting tumor cells, facilitating distant metastases in HER2+ molecular subtype breast tumor models.⁴⁷

Tumor Microenvironment

Tumor cells and the environment surrounding the adjacent neoplastic tissue collaborate to tumor progression. This tumor microenvironment consists of various cell subpopulations, including stromal cells, immune cells, endothelial cells, fibroblasts, and adipocytes, along with hundreds of associated molecules that paradoxically inhibit and induce tumor malignancy.⁴⁸

Endothelial Cells

Tumor-associated endothelial cells develop mechanisms from in situ to invasive stages, including cytokine secretion, receptor activation on tumor cells, and reducing cytotoxic T cell responses.⁴⁹ They also modulate tumor metabolism via paracrine factor interactions. Single-cell RNA-Seq studies by Tirosh et al revealed a link between these cells and Notch signaling through carbonic anhydrase 4 (CA4), highlighting their role, alongside fibroblasts, in regulating angiogenesis within the tumor microenvironment.^50–52

Fibroblasts

Fibroblasts are crucial in homeostasis, involved in mechanical responses, physiological changes, and repair processes.⁵³ In the tumor microenvironment, they promote cancer progression.⁵⁴ Cancer-associated fibroblasts (CAFs) regulate tumor progression and are key drivers of poor prognosis in many cancers, being linked to therapeutic resistance and disease recurrence.⁵⁵ Fibroblast-derived peptide factors suppress immune cells, promoting differentiation, proliferation, and neoplastic growth.⁵⁶ They stimulate pro-inflammatory cytokines like IFN-γ, TNF-α, and IL-1, which suppress T-lymphocyte function.⁵⁷

Adipocytes

Adipocytes primarily influence tumor development in breast and abdominal cavity organ tissues, including the gastrointestinal tract. They contribute to tumor cell proliferation⁵⁸ by supplying metabolic resources, such as long-chain fatty acids, which are catabolized via β-oxidation to increase ATP production, thus enhancing tumor cell proliferation.⁵⁹

Obesity, influenced by both environmental and genetic factors, is a significant risk factor for developing endometrial and breast cancers,⁶⁰ due to adipocyte hypertrophy and hormonal imbalances in these tissues. Cytokine and growth factor-mediated signaling pathways in adipocytes, including Interleukin-6 (IL-6), Tumor Growth Factor Beta (TGF-β), and VEGF, along with increased leptin and decreased adiponectin levels, trigger tumorigenesis and tumor progression.^61,62

Immune System Cells

Innate and adaptive immune cells play dual roles in the tumor microenvironment as tumor-antagonizing (TaICs) or tumor-promoting (TpICs) cells.⁶³ TaICs, including CD8+ and CD4+ T cells, NK cells, macrophages, and neutrophils, combat tumors through cytokine release, apoptosis induction, and the production of pro-inflammatory cytokines.^63–65 Conversely, TpICs, such as regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs), inhibit cytotoxic T cells and promote angiogenesis by producing ECM-degrading metalloproteinases and VEGF.^63–65

The complexity of immune system cells and their tumor-related actions remains poorly understood. Tumor progression is facilitated by evasion mechanisms like the PD-1/PDL-1 axis, where tumor cells express PDL-1 to inhibit cytotoxic T cell activity by binding to T cell PD-1 receptors.^66–69 Additionally, antagonist cells may exhibit paradoxical roles in tumor progression, such as T-cell exhaustion from persistent antigen exposure,^67,70 limited NK-cell activity due to immune checkpoint inhibitors,⁷¹ and the controversial role of B cells, which can influence tumor progression by inducing angiogenesis and inhibiting cytotoxic T cells.⁷²

Emerging Tumor Treatments

Cancer, a complex and heterogeneous group of diseases, has historically posed significant challenges for medical treatment. While tumor resection dates back to antiquity, modern cancer treatment emerged in the 20th century with advancements in surgical techniques and the introduction of radiotherapy and chemotherapy.⁷³ Some discoveries such as the use of mustard gas in tumor cells, identified through cytological analyses showing leukopenia in soldiers during the World Wars, were quite unexpected.⁷⁴

The surge in technological innovation in the late 20th century significantly altered classic tumor treatments. For instance, while conventional radiotherapy is still widely used, four-dimensional image techniques tracking tumor movement and position are now employed to significantly reduce damage to adjacent healthy cells also affected by the treatment.⁷⁵ Currently, advanced therapies, including immunotherapy and Chimeric Antigen Receptor T Cell therapy (CAR-T cell therapy), offer personalized treatment options for various cancer types.

Over the past three decades of biomedical research, more than 100 antibodies have been approved for cancer immunotherapy. These antibodies exert their therapeutic effects through various mechanisms, including: (i) signal pathway inhibition (SPI), (ii) antibody-dependent cell-mediated cytotoxicity (ADCC), (iii) complement-dependent cytotoxicity (CDC), and (iv) antibody-dependent cellular phagocytosis (ADCP).⁷⁶

The main SPI antibodies include cetuximab and panitumumab, which act through competitive inhibition of the Epidermal Growth Factor Receptor (EGFR) and antagonistic action, inducing EGFR internalization, respectively.⁷⁷ Among ADCCs, loncastuximab targets the CD19 antigen primarily expressed on B cells, making it an effective immunotherapy for lymphomas.⁷⁷

In the past decade, immunotherapy has focused on two main pathways of tumor evasion: the PD-1/PDL-1 axis (with Pembrolizumab and Atezolizumab, respectively) and the CTLA-4 immune checkpoint pathway Ipilimumab), the latter being characterized for inhibiting the cytotoxic actions of T lymphocytes by interacting with their ligands (CD80 and CD86) present on antigen-presenting cells.⁷⁷

CAR-T Cells in Cancer Treatment

CAR-T Cell Therapy has emerged as a paradigm-shifting revolution in Oncology, offering a novel approach to treating refractory tumors. By genetically reprogramming a patient’s T cells, this therapy provides a targeted immune response against malignant cells, radically changing the therapeutic landscape. Its technological foundation involves producing chimeric antigens linked to T cells, typically through transgenic insertion assays.⁷⁸ These synthetic receptors must conform to a rigid structural template, consisting of four main components: an extracellular antigen-binding domain, a hinge region, a transmembrane domain, and intracellular signaling domains.^79,80

The cornerstone of CAR-T cell therapy lies in the specific interaction of chimeric antigens with proteins predominantly or exclusively expressed in tumors, rendering this therapy one of the most prominent in recent years.

In a constantly developing field, to this date, the Food and Drug Administration (FDA) has approved six CAR-T cell treatments in the USA,⁸¹ namely:

• Kymriah™ (tisagenlecleucel): for patients with acute lymphoblastic leukemia, B-cell lymphomas, and follicular lymphomas, featuring a chimeric antigen receptor for CD19;
• Yescarta™ (axicabtagene ciloleucel): for patients with B-cell lymphomas and follicular lymphomas, featuring a chimeric antigen receptor for CD19;
• Tecartus™ (brexucabtagene autoleucal): for patients with mantle-cell lymphoma and acute lymphoblastic leukemia, featuring a chimeric antigen receptor for CD19;
• Breyanzi™ (lisocabtagene maraleucel): for patients with B-cell lymphomas, featuring a chimeric antigen receptor for CD19;
• Abecma™ (idecabtagene vicleucel): for patients with multiple myeloma, featuring a chimeric antigen receptor for BCMA;
• Carvykti™ (citacabtagene autoleucel): for patients with multiple myeloma, featuring a chimeric antigen receptor for BCMA.

Extracellular Vesicles and Tumors

Extracellular Vesicles (EVs), involved in autocrine, paracrine, and endocrine signaling processes, are crucial for embryonic developmental stages, influencing cell differentiation and proliferation via Wnt and Hedgehog protein gradients,⁸² neural crosstalk,⁸³ and immunological signaling, activating Antigen Presenting Cells (APCs) and T lymphocytes.⁸⁴ Given their association with cell proliferation, migration, differentiation, and immune cell activation, EVs have been closely linked to tumorigenesis and tumor progression. Moreover, the discovery of EVs in liquid tumor biopsies has solidified their role as critical biomarkers. Substantial differences in surface proteins and intracellular molecules such as miRNAs between early and advanced-stage tumors render EVs an invaluable tool for cancer prognosis.⁸⁵

Tumor cells release significantly more EVs than normal cells, highlighting their crucial role in tumor-microenvironment communication and subsequent malignant progression.⁸⁶ This EV-mediated communication is involved in angiogenesis induction, release of pro-metastatic factors and immune suppression, notably through actions of siRNAs,⁸⁷ induction of chemokine secretion by macrophages and synthesis of TGFβ,⁸⁸ in addition to EMT through EVs released by tumor cells under hypoxic conditions.⁸⁹

Additionally, EVs actively contribute to ECM remodeling within the tumor microenvironment. By transferring proteins such as fibronectin, annexins, and integrins to the tumor stroma via exosomes, EVs facilitate tumor invasion through the formation of new ECM scaffolds.^90,91

Over the past two decades, EV research has significantly advanced, leading to their dual application in cancer therapy. EVs serve not only as therapeutic targets but also as versatile drug delivery vehicles capable of directly targeting tumor cells and the microenvironment.⁴¹ Their small size, versatile cell-entry mechanisms, biocompatibility and potential signal peptide modification in their structures render EVs ideal candidates for drug delivery systems.⁹²

Pan-Cancer Research

Despite the numerous cancer divergencies and its establishment as a heterogeneous set of pathologies, the landscape emerging from a significant number of convergences is as rich and complex as its individualized studies.

Over the past two decades, extensive research has shaped the field of pan-cancer analysis; however, the publication of “Pan-cancer analysis of whole genomes” initiative, published in Nature was indeed a pivotal moment for the field.⁹³ This publication sparked increased research in the field, coinciding with similar efforts by Cell Press through TCGA Consortium, funded by the National Cancer Institute and the National Human Genome Research Institute. These efforts aim to comprehensively analyze cancer by examining tumor cells, their molecular similarities across different tumor types, associated oncological processes involved signaling pathways, and the genetic and epigenetic changes driving human tumor development.

Tumor Cell-of-Origin Patterns

Cell-of-origin refers to the genetic changes a normal cell undergoes as it becomes cancerous. Understanding these alterations is crucial for pan-cancer analysis, identifying clusters with associated modifications across tumors with respect to aneuploidies, DNA hypermethylation, transcriptional expression of coding and non-coding genes, microRNA transcription, and protein expression.⁹⁴ Hoadley et al, in their analysis of 10,000 tumor samples, highlighted a characteristic heterogeneity regarding aneuploidy, dividing 33 types of tumors into 10 different clusters, particularly noting those with a higher aneuploidy index. Regarding epigenomic changes, over 3000 tumor samples exhibited hypermethylation in CpG islands, forming 25 clusters with similarities in methylation positions; therefore, the epigenomic mechanism is directly associated with similar cellular populations, governed by pre-existing transcription programs.

In turn, mRNA analyses revealed insights into gene expression linked to tumorigenesis. Similar morphology cells showed similar transcriptional potential compared to cells of more distinct morphologies and origins. The results for microRNAs were more heterogeneous, indicating a more effective integration among more distantly related tumors.⁹⁴

By linking oncogenic processes, researchers can more precisely identify specific changes across different tumor types and search for antitumoral drugs for oncological therapy. In 2018, using TCGA data, Sanchez-Vega et al conducted a meticulous analysis that established mechanistic relationships between ten classic pathways studied in oncology.⁹⁵ Their findings included:

• A high average RTK-RAS signaling pathway alterations across all analyzed tumor types, totaling 46% of the studied samples;
• KRAS is the most frequently altered gene (9% of the samples);
• BRAF gene alterations were found in over 50% of the melanoma and thyroid carcinoma samples;
• Synergistic relationships between different pathways in different tumor types directly influence each other’s changes.

Thorsson et al shed light on the immune system’s role in tumor initiation. By extensively analyzing TCGA data across 33 cancer types, Thorsson et al identified six distinct immune response patterns within the neoplastic environment, namely: wound healing, IFN-γ dominant, inflammation, lymphocyte depletion, immune silencing and TGF-β dominant. These classifications were based on analyzing markers of innate and adaptive immune systems and exploring the immune population heterogeneity in different tumor types, the presence of aneuploidies and gene expression and co-expression of modulatory genes.

The data suggest a strong link between immune responses and patient survival across different tumor types. Additionally, the study revealed how immune evasion varies across different tumor-associated microenvironments. Intriguingly, tumors with dominant immune responses displayed a wider variety of B and T cell lymphocyte receptors (TCRs and BCRs), leading to a more effective response against the tumor environment. In contrast, other tumor types favored immune escape and tissue inflammation.⁹⁶

Based on data from the Pan-Cancer Analysis of Whole Genomes initiative, the results have increasingly correlated somatic mutations as significant factors in tumor initiation. The analyses revealed over 43 million somatic alterations, 2.4 million somatic indels, and 19.4 million retrotransposition events across the data.⁹³

Among the findings, key points of interest in biomedical and clinical research included:

• Dual inactivation of TP53 alleles triggers instability in tumor suppressor genes;
• Three processes are associated with genomic reconfiguration in tumor emergence: chromoplexy, kataegis, and chromothripsis.

Chromoplexy, a process of inappropriate DNA reorganization through repair mechanisms among different chromosomes, was identified in over 17% of the samples, especially in prostatic adenocarcinomas and lymphoid tumors. Kataegis, characterized by defined sites of gene hypermutations, was found in over 60% of tumors, notably in lung carcinomas, bladder tumors, melanomas, and sarcomas. Chromothripsis, defined by extensive fragmentation of a chromosome into small fragments, followed by random reorganization and reconnection, occurred in 22% of the total samples, with a higher association with sarcomas, glioblastoma, and breast adenocarcinoma.⁹³

Pan-cancer research also extensively studies the tumor microenvironment. Its heterogeneity poses a significant challenge, addressed through the use of single-cell RNA-Seq, in a remarkable effort by Nofech-Mozes et al. By analyzing over 300,000 tumor microenvironment-related cells across 19 common tumor types, using comparative machine learning, researchers extracted gene expression data from tumor cells and surrounding tissues. This allowed them to create cell/tumor type maps identifying the tissue origin of tumors and understanding how cells interact within tumors.⁹⁷

In short, pan-cancer research is essential due to the complex and varied nature of tumors, with numerous regulatory and modulatory mechanisms. Therefore, by studying multiple cancer types together, researchers can identify commonalities and differences, leading to improved diagnosis, prognosis, and treatment. This approach holds the promise of more accurate and personalized therapies, which is crucial for a medium- to long-term reduction in global mortality rates due to different cancers.

Results and Discussion

We undertook a global analysis of epidemiological, genetic, and protein data across 19 cancer types using three large public repositories, namely: the Pan-cancer analysis of Whole Genomes (ICGC/TCGA), China Pan-cancer, and Metastatic Solid Cancers. These datasets combined include over 13,616 tumor samples, with 21% representing advanced (stage IV), or metastatic tumors (Figure 2A and B).

Figure 2 Datasets of Pan-cancer. (A) Pan-cancer repositories used for in silico analyses; (B) Samples divided by tumor stage; (C and D) Correlations of sample and patient numbers and primary tumor site, metastasis, recurrence and site of recurrence; (E) Samples divided by gender.

A standardized patient profile was established among the samples, gathering information on cell type, recurrence site and whether the tumor was primary or metastatic. By analyzing patient records, which could include multiple samples per patient, the dataset contained more samples and patients with primary tumors, followed by metastatic tumors (Figure 2C and D).

By overlapping patients, a correlation is observed between tumor recurrence and the original and subsequent tumor locations. Recurrences often appeared at the initial treatment site (eg, breast cancer returning in the breast) or in tissues similar to the original tumor (eg, breast cancer recurring in lymph node-rich areas or reproductive organs, such as salivary gland and ovarian tumors).

Regardless of age and ethnicity, gender analysis revealed that women are more likely to develop cancer than men (six out of 10), excluding non-melanoma skin cancer. Specifically, breast cancer is the most common in women, while lung cancer (excluding melanomas) is the most prevalent in men. These findings also align with smoking rates, wherein 37% of smokers worldwide are men, and only 7% are women, highlighting a correlation with lung tumors.⁹⁸

Results on ethnic epidemiology and mutated gene frequency provide insights into the convergence of tumors worldwide. The top 10 genes with the highest alteration frequency across all 20 tumor types occur in all global populations. However, the BICRA gene, involved in chromatin remodeling and proposed as a glioma suppressor gene, shows a higher mutation frequency in South Asian populations. This observation requires more in-depth longitudinal clinical analyses to establish a relationship between the incidence of this tumor type in that population.

An analysis of the main types of genetic alterations across the 20 tumor types studied here reveals intriguing patterns in tumor development and progression. The association of ADAM family genes with ECM reorganization is particularly enriched in pancreatic tumors. Such a correlation links the progression of this tumor type to cell adhesion dysregulation and potential aggressiveness, primarily through metastasis, as well as a morphological remodeling of tumor cells, potentially through a more effective EMT mechanism.

Protein interaction predictions were conducted among the 48 most mutated genes in the tumor types studied here (Table 5). Within the analyzed sample size (>13,000), 40% of these genes are not categorized as tumor-associated genes, highlighting the importance of bioassays to correlate them with tumorigenesis and tumor progression.

Table 5 Frequency of the Main Mutated Genes in Tumors in 13777 Samples

Protein interaction pathways were clustered using the k-means algorithm, an unsupervised learning method, to group proteins based on family similarity and predicted interactions from in silico analyses (Table 6 and Figure 3A–E). Based on these predictions and established pathways among proteins within the same clusters or through protein interactions across different clusters, the following hypotheses can be postulated:

• CTNNB1 possibly modulates the cell cycle by alterations in the Wnt/Beta-Catenin signaling pathway via APC and activating the MAPK pathway and others involved with proteins encoded by the KRAS and RB1 genes, thus relating to uncontrolled proliferation processes. It also directly acts on negative feedback mechanisms in apoptotic processes, in addition to the positive feedback of cell migration processes via the expression of atypical cadherins (FAT1);
• KMT family proteins (such as KMT2A, KMT2B, KMT2C, KMT2D and KMT2E) are mainly responsible for tumorigenesis processes via epigenetic modulation by methylation of histones at lysine residues;
• RB1 is closely linked to the regulation of tumorigenic processes in glioblastomas through interactions with proliferation mechanisms involving alterations in genes such as TP53, KRAS, and EGFR;
• HMCN1 may be involved in increased tumor invasion through modulation of tumor-associated fibroblasts.

Table 6 Protein Clusters and Main Associated Mechanisms

Figure 3 Protein predictions networks by StringDB Analysis. Protein prediction analyses via StringDB, delimiting the (A) global interactions between the main protein-coding genes mutated in the pan-cancer analyses and Ontological analysis of Biological Processes (main cluster) and (B–E) local interactions, through subclusters 1–4, respectively, with data on the mutational frequency of the genes of each subcluster in the 19 types of tumors analyzed. The protein interactions evidenced in the clusters indicate relationships evidenced both experimentally (Purple line) and predicted in silico (Red line - presence of fusion evidence; Green line - neighborhood evidence; Blue line - cooccurrence evidence; Yellow line - textmining evidence; Light blue line - database evidence; and Black line - coexpression evidence). The interactions between subclusters are seen through the dashed lines.

Gender-dependent enrichment analyses were also performed and showed that the main genes correlated with tumorigenesis have similar expression in men and women, but genes such as DNAH5, ADGRV1 (in women) and EYS (in men) appear mutated more frequently (Figure 4A–D).

Figure 4 Frequency of gene mutations of the main genes of the protein interaction subclusters by gender. (A–D) Mutational frequency of genes in each subcluster of male (blue bars) and female (pink bars) genders, according to the analyzed subclusters 1–4.

Ultimately, tumor convergence, beyond the search for novel proteins or associated molecules, is rooted in well-known genes involved in tumoral processes. However, the study of these genes should not only focus on proliferation pathways and the various already well-understood processes but also on the relationships between different genes not directly related to tumors. Various genes have interactions in tumors that are distinct in terms of the main pathways associated with their malignancy, progression and others secondary processes, as shown in Figure 5:

Figure 5 Tumor hallmarks associated with the analyzed genes. Association between the main tumor hallmarks and the most frequently mutated genes in the pan-cancer analyses performed.

The relationship between these processes and the genes involved and delimited by pan-cancer analyses demonstrates the intricate gene network involved in the processes of tumorigenesis and tumor progression, bringing light to new molecules that can help in the landscape of basic, applied and clinical oncology research.

These findings underscore the necessity of considering genes previously unrecognized in the tumor context yet exhibiting altered expression across various cancer types. While the heterogeneity of available data poses challenges, it also provides opportunities to uncover new molecular interactions and genetic networks pivotal to tumorigenesis. Integrative bioinformatics approaches can identify innovative therapeutic targets and prognostic biomarkers, enhancing our understanding of the underlying mechanisms of cancer. Therefore, a critical and thorough analysis of these data is essential to reveal hidden biological potential and facilitate significant advancements in oncology, ultimately improving diagnostics, prognostics, and personalized therapies for patients.

Conclusions

Our literature review reveals a complex network of relationships among different tumor types, particularly in terms of molecular interactions and pathways activation, triggering uncontrolled proliferation, migration, and invasion mechanisms. Studies have mainly focused on the convergences that different molecular types and subtypes share within the context of pan-cancer.

Convergence patterns have also been identified and highlighted in seminal works, with significant additions from multi-omics data analyses, demonstrating the existence of diverse regulatory networks and protein interactions with potential clinical interest in the future.

Comparative analyses of different tumor types provided a comprehensive overview of molecular similarities and disparities, enabling the identification of potential therapeutic targets applicable across multiple oncological scenarios, such as those involving CTNNB1, APC, KMT2D and RB1 genes. This integrated approach underscores the importance of translational research and precision medicine for a better understanding of the shared and distinct molecular basis across various cancer types.

In conclusion, our findings underscore the intricate interplay between epidemiological, genetic and molecular factors driving tumorigenesis, offering a holistic view of pan-cancer oncology. This profound knowledge is essential for designing more targeted and effective therapeutic interventions, ultimately leading to improved control and treatment of the diverse spectrum of cancers.