The TP53 tumor suppressor gene: From molecular biology to clinical investigations

Abstract

Extensively studied over the past four decades, the TP53 gene has emerged as a pivotal watchman in cellular defense and a key factor in cancer biology. TP53 is the most frequently mutated gene in human malignancies, 50% of which carry alterations to it. Initially, the functions of p53 were thought to be restricted to cell‐cycle arrest and apoptosis. With time, however, a growing number of functions have been discovered, illustrating p53's role as a master switch between any cellular stress and cellular or multicellular responses that contribute to its anti‐tumor activity. Indeed, the peculiar landscape of TP53 mutations and its high heterogeneity are linked both to the structure of the protein and its ubiquitous function in regulating cellular homeostasis. Mutations in p53 are associated with poor response to therapy and shorter survival in most cancer types, and the diagnosis of p53 mutations is currently used to improve case management in some types of leukemia and lymphoma. Although TP53 has been defined as a tumor suppressor gene, overexpressed mutated p53 variants found in human tumors are defined as dominant oncogenes with a potential gain of function, which makes the gene a very attractive target for developing new cancer treatments. Beyond its role in cancer, our review also highlights TP53's significance in non‐neoplastic conditions, such as bone marrow failure syndromes and certain developmental disorders, where chronic p53 activation plays a crucial role in cellular stress responses, demonstrating its broader biological importance.

Introduction

Over the years, the p53 gene (TP53 hereafter) has earned such monikers as the “guardian of the genome” [1], the “death star” [2], the “good and bad cop” [3], an “acrobat in tumorigenesis” [4], or the “unlucky gene” [5], among others. TP53 has inspired the publication of more than 120,000 articles, 13,000 reviews, and 500 meta‐analyses; the sequencing of several hundred thousand tumors; and the description of 250,000 mutations. From this enormous body of data, what relevant information can be surfaced for clinicians? The first indisputable fact is the high frequency of TP53 alterations in human cancers [6], generally manifesting as genetic modifications in the gene's coding region and resulting in the expression of an inactive p53 protein. Additionally, in tumors associated with viral infection (e.g., cervical cancer or certain head and neck cancers caused by human papilloma viruses (HPV)), the virally expressed E6 protein induces cellular degradation of p53, thereby compromising its function. Similarly, in osteosarcoma and some brain tumors, amplification of MDM2, the major negative regulator of p53, leads to increased instability of the protein, akin to HPV infection. The frequency of TP53 alteration is highly heterogeneous across various cancer types: as little as a few percent in thyroid cancer, but up to 30% in breast cancer, and nearly 100% in high‐grade serous ovarian cancer or small cell lung cancer (Fig. 1a). In several cancer types (prostate and leukemia), the frequency of TP53 mutations at diagnosis can be low, but it may increase upon disease progression or during therapy. All information considered, an estimated 50% of human cancers carry a TP53 alteration, making this gene the most frequently mutated in human malignancies [6, 7].

Fig. 1.

TP53 mutation in human cancer: (a) worldwide cancer burden (new cases and number of deaths, Globocan 2022) and average TP53 mutation frequency in the 14 most frequent cancers. * In cervical cancer, loss of TP53 function occurs by interaction with the E6 protein of oncogenic human papilloma viruses. (b) Schema showing the domain structure of the p53 protein, including the transactivation domains (TAD 1 and 2), the proline‐rich domain (PRD), the DNA‐binding domain (DBD), the tetramerization domain (TET), and the negative regulatory domain (NEG‐REG). The aligned graphs indicate the relative frequency of mutations across different domains of p53. The DNA‐binding domain is the most frequent location for p53 mutations, according to the UMD_TP53 database. High‐penetrant p53 variants can be either structural or contact variants, depending on the function of the residue. In contrast, residue association with p53 multimerization shows a lower penetrance with only a partial activity loss. (c) X‐ray structure of p53 DBD showing locations of the most frequent mutations. DNA contact variants at positions 280, 248, or 273 are localized in the vicinity of the DNA molecules, whereas structural variants (175 or 220) are buried inside the molecule (taken from [8]).

Various murine models demonstrate the significance of p53 in tumor suppression. Mice homozygous for TP53 gene deletion are viable and fertile but exhibit a high susceptibility to cancer, with all p53‐null mice succumbing to the disease by 10 months of age [9, 10]. Heterozygous mice are also at risk of developing cancer but tend to die at older ages. Mice with different TP53 variants, akin to those observed in human cancers, develop a more diverse range of tumor types depending on the specific variant expressed and the genetic background of the mice [11, 12, 13].

In the first part of this review, we will focus on the p53 protein and how it regulates various cellular pathways. Indeed, the peculiar landscape of TP53 mutations and the high heterogeneity of the latter are linked to the structure of the protein as well as its ubiquitous function in the regulation of cellular homeostasis, two features that took more than 30 years of research to unravel. In the second part, we will discuss the clinical value of TP53 research and the gene's potential for use as a cancer biomarker.

The field of investigation regarding p53 is quite broad, and we will not be able to cover all of it here. The literature does, however, provide recent reviews focused on TP53 isoforms [14] and articles focused on other members of the p53 family (p63 and p73) [15], as well as various p53‐based therapeutic strategies [16, 17] and the detailed analysis of the multiple pathways regulated by p53 [18, 19, 20].

The p53 protein: structure and function

Cancer‐related genes are primarily classified into loss‐of‐function tumor suppressor genes (TSGs) and hyperactive oncogenes. However, that perspective is oversimplified and requires reevaluation, as genes involved in DNA replication, repair, or immune evasion have also been identified as significant contributors to cancer [21, 22, 23, 24]. Activation of oncogenes usually occurs through missense variants within the protein's functional domain, clustering at specific sites. Most of these changes are heterozygous, preserving one intact allele. In contrast, loss‐of‐function alterations, such as indels, splicing, or nonsense mutations, in TSGs prevent protein expression and are spread throughout the protein sequence. For many TSGs, both alleles must be compromised. TP53 is a negative regulator of cellular proliferation and is classed as a TSG, but it does not conform to this pattern; most mutational events in it are missense mutations dispersed across more than half of the protein (Fig. 1b,c) [6].

The p53 protein is a transcription factor comprising five domains: a transactivation domain divided into two regions responsible for activating the transcription of multiple genes, a proline‐rich domain (PRD) involved in the regulation of apoptosis, a 200‐amino‐acid‐residue DNA‐binding domain (DBD), a tetramerization domain (TET) that enables the oligomerization of the p53 monomers, and a negative regulatory domain (NEG‐REG) containing multiple sites for post‐translational modifications (PTMs) involved in p53 stability (Fig. 1b) [25, 26].

In normal cells, p53 is expressed at very low, often undetectable, levels despite discernible mRNA expression. The primary reason for this is that p53 is targeted for proteasomal degradation by the E3 ubiquitin ligase MDM2, assisted by MDMX (Fig. 2). In these normal conditions, p53 mainly exists as a cytoplasmic dimer with low DNA binding efficiency [26, 27]. When the cell experiences stress (such as DNA damage, hypoxia, or oncogene activation), various pathways impair the interaction between MDM2 and p53, resulting in p53 accumulation. The protein then shifts to a tetrameric form with enhanced DNA binding activity and localizes to the cell nucleus. Once activated, p53 directly initiates transcriptional programs that regulate cellular stress responses (Fig. 2). PTMs are crucial for modulating p53's stability, localization, and transcriptional activity in response to cellular stress signals. Phosphorylation—primarily mediated by kinases such as ATM and ATR—enhances p53 activation by inhibiting its degradation via MDM2, whereas acetylation by histone acetyltransferases such as p300/CBP increases its ability to bind DNA and promotes the expression of target genes involved in cell cycle arrest and apoptosis. Conversely, ubiquitination—mainly driven by MDM2—marks p53 for proteasomal degradation, ensuring tight regulation of its activity under normal circumstances. Furthermore, other PTMs—such as methylation, sumoylation, and glycosylation—further diversify p53's functional outcomes, influencing its capacity to suppress tumor development [28].

Fig. 2.

p53 signaling pathways. Without stress, the MDM2/MDMx complex ubiquitinates p53, leading to its subsequent destruction by the 26S proteasome. Upon various types of stress, multiple pathways prevent p53 degradation and activate p53, which will accumulate in the cell nucleus to orchestrate an adequate transcriptional response according to the type of stress.

The TET domain (residues 326–356)—necessary for the formation of p53 tetramers—exhibits an unusual structural organization, forming a dimer‐of‐dimers in which two dimers interact to create a stable tetramer. Each dimer is stabilized by an antiparallel β‐sheet and an α‐helix, facilitating strong cooperative binding that increases p53's affinity for DNA and its regulatory functions [29, 30]. This unique arrangement ensures structural flexibility and functional adaptability, allowing p53 to effectively regulate gene expression in response to cellular stress. Residues 180 and 181, located in the DBD, are also essential for p53 dimerization [31].

The p53 NEG‐REG is intrinsically disordered and can fold back onto the central DBD, preventing p53 from efficiently binding to its target DNA sequence. This ensures that p53 remains inactive under normal cellular conditions and only binds DNA when necessary.

The p53 DBD is unusually long, comprising 600 coding nucleotides across 5 exons and 200 residues. It includes multiple loops and sheets, with amino acid residues either in direct contact with DNA (K120, S241, R248, R273, A276, A277, R280, R283), involved in the folding of the domain to ensure proper structure (R175, R249), or associated with zinc chelation, an essential feature for p53 stability and DNA binding (R176, H179, H238, C242) (Fig. 1b,c). Systematic functional and structural studies have shown that the p53 DBD is extremely fragile, as most single amino acid substitutions located within this domain impair TP53 functionality. This extreme vulnerability of the domain gives rise to the particular landscape of TP53 gene mutations in human cancer [5]. Various saturation mutagenesis studies, with substitutions at every position in this domain, have demonstrated that these changes lead to partial or total loss of DNA binding activity [32, 33, 34, 35].

Notably, the sequence of the p53 response element (p53‐RE) elucidates the high diversity of pathways regulated by the protein. P53‐REs consist of 2 10‐base pair (bp) palindromic DNA segments with the consensus sequence of 5′‐Pu1Pu2Pu3C4(A/T)5(A/T)5′G4′Py3′Py2′Py1′‐3′ for each of the two half‐sites, where Pu and Py denote purine and pyrimidine bases, respectively [36, 37]. Any two half‐sites are usually separated by 1–3 bps, but some may be separated by as many as 20 bps. Hundreds of p53‐REs have been identified in the regulatory regions of p53 target genes. However, their high degeneracy is linked with varying affinities of p53 and, consequently, different levels in transcriptional response. This could be a key feature of the significant diversity of p53 responses to various types of stress [38].

TP53 variant heterogeneity

The functional and structural heterogeneity of p53 missense variants found in human cancers was first recognized over 30 years ago with the identification of the so‐called structural mutants, in which the DBD is unfolded (e.g., p.R175H), and the “DNA contact” mutants, in which residues that interact directly with DNA are substituted (e.g., p.R273H) [39, 40]. Nuclear magnetic resonance, circular dichroism, and x‐ray diffraction analyses have confirmed and expanded this description, suggesting that multiple thermodynamic stages are associated with the various mutations depending on their position [41]. This heterogeneity also extends to the biological activity of the majority of mutant p53 proteins, as demonstrated by the various multiplex assays of variant effect conducted on over 10,000 p53 variants [32, 33, 34, 35]. Furthermore, depending on the location of the substitution (residues involved in DNA contact or in protein stability), the dysfunctional protein will exhibit different characteristics, ranging from a simple loss of DNA binding activity to complete denaturation, resulting in heterogeneous penetrance in tumors [42]. The analysis of p53 variants in codon 175 serves as a prime example of this heterogeneity. Variant p.R175H is fully denatured and displays a total loss of function (LOF), including impaired growth arrest and apoptosis [39, 43]. In contrast, variant p.R175P, also found in human tumors, demonstrates normal cell cycle arrest and gene p21 induction behavior but shows deficiency for apoptotic activity and fails to transactivate apoptotic genes. Mice homozygous for the p.R172P mutation (equivalent to the human p.R175P alteration) exhibit defects in TP53‐dependent apoptosis but retain a partial cell cycle checkpoint function [44]. These mice show a significantly lower predisposition to developing tumors compared to mice lacking TP53 expression. Additionally, these tumors do not exhibit the chromosomal instability observed in TP53 or in mice expressing the hot‐spot mutant R175H. Variants at codon 181 also exemplify the heterogeneity of the penetrance of TP53 mutations. Variant p.R181P is frequently found as a somatic mutation in human tumors and is associated with a total loss of activity. In contrast, variant p.R181C is much less frequently observed as a somatic event and demonstrates only a slight loss of activity [42]. Furthermore, variant p.R181C has been identified as a causal germline variant in families associated with an increased risk of cancer, albeit with low penetrance [45, 46].

Although wild‐type p53 is unstable due to its association with MDM2, most missense p53 variants observed in tumors are highly stable and accumulate in the cell nucleus [47]. Immunohistochemical staining for p53 as a surrogate for mutation analysis has a long history of use, but it has fallen by the wayside in favor of Sanger sequencing and now next‐generation sequencing (NGS), as it also detects nonsense and frameshift mutations, which are not expressed [48]. This accumulation of p53 in tumor cells has led to the hypothesis that beyond a LOF, mutant p53 could behave like a dominant oncogene via two different but non‐exclusive mechanisms.

The first mechanism is a dominant‐negative effect (DNE) in cells that express both wild‐type and mutant alleles [49]. Mixed p53 tetramers, with both wild‐type and mutant p53, are significantly less efficient at activating the wild‐type p53 target genes than tetramers containing only wild‐type p53. Variants in the TET domain that cannot undergo dimerization or tetramerization do not display any DNE. Although DNE is largely undisputed, its clinical relevance remains to be established. As in most cancers, the p53 wild‐type allele is lost, and it is unlikely that DNE is sufficient to abrogate all p53 function [19, 50, 51].

The second mechanism is a gain of function (GOF) with mutant p53 acquiring new oncogenic activity [52]. Multiple studies, using either cellular models expressing specific p53 variants or mouse models, suggest an oncogenic GOF for specific p53 variants via the activation of specific pathways associated with metastasis, metabolic reprogramming, or cell differentiation [53]. Of note, though, more recent studies using more relevant isogenic models have challenged this GOF hypothesis, suggesting that LOF and DNE are the most important features of mutant p53 [54].

TP53 pathways

Historically, CDKN1A—encoding the cyclin dependent kinase (CDK) inhibitor 1A protein p21Cip1 (p21 hereafter)—was the first gene shown to be positively regulated by p53 [55]. Following genotoxic stress, activation of the ATM kinase impairs p53‐MDM2 interaction by phosphorylating both proteins. This leads to p21 expression and cell cycle arrest via the p21‐mediated inhibition of cyclin/CDK complexes (Fig. 3a). This inactivation of the cyclinE/A‐CDK2 and cyclinD‐CDK4/6 complexes inhibits CDK‐mediated phosphorylation of pRB, preventing the release of E2F transcription factors, which control G1/S transition. This growth arrest in G1 prevents the replication of damaged DNA. The p21 promoter has been shown to contain two p53‐REs with a strong affinity for p53 [56].

Fig. 3.

The dual function of p21 in response to p53 activation. (a) In the “classical” pathway, cell cycle progression requires hypophosphorylation of RB to prevent binding to E2F transcription factors and form a repressive complex that block the expression of genes required for DNA replication and S‐phase entry. Expression of p21 prevents RB phosphorylation, leaving RB complexed with E2F factors and histone deacetylases (HDAC) to maintain a repressive chromatin state and block S‐phase gene expression. (b) In the p53‐p21 DREAM pathway, multiple genes associated with cell cycle regulation or DNA repair are indirectly negatively regulated by p53 via RB hypophosphorylation. CDK, cyclin dependent kinase.

TP53 also controls the expression of some of the key initiators of apoptotic cell death induced by diverse stresses, such as nutrient deprivation or DNA damage [57]. Two pathways are considered the major drivers of apoptosis in all cells: the intrinsic pathway, initiated by the formation of BAX and BAK pores on the outer mitochondrial membrane, and the extrinsic pathway, triggered by death receptors on the plasma membrane. Both pathways are regulated by p53 [58].

In the intrinsic pathway, BAX, PUMA, and NOXA are directly transcribed by p53, and the two latter are essential for apoptosis in hematopoietic cells. In the extrinsic pathway, the death receptors FAS (CD95/APO‐1) and DR5 (TRAIL‐R2) are also p53 target genes. The importance of apoptosis in regulating cell death during normal tissue development and neoplastic transformation, and as an essential response to chemotherapeutic agents, has led to the general belief that it is the main cellular response driving the p53 tumor‐suppressive effect. Although this could be true for some specific cases, a range of work in mouse models has shown that tumor regression via p53 overexpression uses different pathways [59, 60]. Indeed, mice deficient in p21, PUMA, and NOXA or deficient in targeting most p53 target genes are not prone to cancer [61, 62, 63].

Subsequent studies have unraveled several layers of complexity within the p53 pathways. First, due to the high degeneracy of the p53‐RE, several hundred genes have been shown to be activated by p53 and to carry p53‐RE in their regulatory regions [64]. These genes are associated with multiple cellular outcomes beyond growth arrest or apoptosis, including DNA repair, autophagy, or senescence, among others (Fig. 2). Second, most types of cellular stress are able to induce a p53 response, but they all use different mechanisms to activate the protein [65, 66, 67]. Therefore, depending on the type and extent of the stress or the cell type, p53 responses will vary widely. For example, lymphocytes are very prone to dying via p53‐dependent apoptosis after DNA damage, whereas colorectal or liver cells sustain growth arrest or enter a senescence state. More recent studies have shown that p53 can also modulate innate and adaptive immunity involved in the homeostatic regulation of immune responses [68, 69, 70].

Large‐scale transcriptomic studies have confirmed that wild‐type p53 can activate multiple genes, but other work has shown that it can also act as a transcriptional repressor [71, 72]. This previously underestimated activity of p53 becomes more pertinent with the observation that many of these repressed genes are deregulated in human cancer expressing mutant p53.

This transcriptional repression is indirect because p53 does not bind to regulatory elements in the concerned genes. Still, it requires p21, which is a key element in the p53‐p21 DREAM pathway [71, 73, 74]. DREAM (dimerization partner, RB‐like, E2F, and multi‐vulval class B) is a multi‐protein complex localized in the CHR box of the promoters of hundreds of cell‐cycle‐regulated genes associated with the regulation of cellular checkpoints, DNA replication, or DNA repair [75]. Upon p21 expression in response to p53, hypophosphorylated Rb will not be released from this repressive complex, thus ensuring that cells do not progress through the cell cycle, whereas damaged (Fig. 3b). In tumors with p53 loss, dysregulation of the p21‐DREAM pathway will lead to uncontrolled proliferation and survival of damaged cells.

TP53 mutation in cancer

TP53 is located on the short arm of chromosome 17 (17p), and the deletion of this latter was found to be a common feature in human cancer as early as 1985 [76]. The development of PCR‐associated molecular diagnostics empowered the description of the first p53 mutations in colon and lung cancer in 1989, and such mutations were rapidly found to be common in other types of cancer [76, 77, 78]. Today, NGS (also called massively parallel sequencing) has enabled such sequencing projects as The Cancer Genome Atlas (TCGA) or The International Cancer Genome Consortium (ICGC). These projects are aimed at sequencing the genomes of primary cancer samples (11,000 and 25,000 cases for TCGA and ICGC, respectively) and have provided a clearer picture of the TP53 mutation landscape in human cancer. Their results earned TP53 the now infamous title of “the most frequent mutated gene in human cancer” (Fig. 1) [79, 80]. Because it is a TSG, both alleles of TP53 need to be inactivated. As observed in multiple cancer types, this manifests classically and most frequently as a mutation in one allele followed by a loss of the second allele via a large or full deletion of the short arm of chromosome 17. Copy‐neutral losses of heterozygosity (deletion of the wild‐type allele and its replacement by a copy of the mutated allele) have been observed in several situations (mainly hematological malignancies) and may be underestimated due to their complexity. Situations involving different mutations in the two different alleles have also been observed, but they are very infrequent [6].

High expectations regarding the clinical utility of TP53 alteration as a novel biomarker drove a tremendous amount of research focused on understanding the tumor suppressive effect of p53. Unfortunately, to date, that promise for clinical applications has not been met [17, 81, 82]. Cancer biomarkers play a crucial role in outlining the prognosis of a disease independently of any treatment (known as prognostic biomarkers) or in predicting how a cancer will respond to a specific treatment (predictive biomarkers).

Several tens of thousands of studies have analyzed the impact of TP53 mutations on survival or response to therapy in different types of cancer. Most point to associations between p53 loss of activity in tumors with genetic instability, poor responses to therapy, and shorter patient survival (Fig. 4). However, any clinical applications of these findings are, so far, restricted to just a few cancer types, such as chronic lymphocytic leukemia (CLL) or acute myeloid leukemia (AML) [83, 84].

Fig. 4.

Deregulation of the p53 network in cancer cells. Dysfunctional p53 will not bind to the response element localized in the promoters of all p53‐regulated genes, including p21. Impairing this transcriptional pathway (left part of the Fig.) prevents the activation of the various p53 response genes. Lack of p21 induction also impairs the DREAM pathway (right part of the Fig.), leaving multiple genes repressed by the DREAM complex, as shown in Fig. 3.

TP53 mutations have been described in CLL since the early 1990s, and an association between TP53 mutations, drug resistance, and poor clinical outcomes was first demonstrated in 1993 [85]. Then, in 2000, a study using FISH analysis for multiple chromosomal markers showed that 17p deletion was an independent predictor of disease progression and survival [86]. Subsequent works using either 17p deletion or TP53 mutation as a biomarker confirmed this finding and resulted in TP53’s classification as a well‐established prognostic marker associated with refractory responses to chemotherapy. The introduction of novel targeted agents has greatly altered the clinical course of CLL patients, who now benefit from responses never observed during the chemoimmunotherapy era. The detection of del(17p) and TP53 mutations has now become an integral part of routine diagnostics, as recommended by the TP53 network of the European Research Initiative on CLL [87].

AML is a deadly hematological malignancy that occurs most commonly (but not exclusively) in the elderly population (detected in more than 50% of individuals aged ≥65 years) [88]. Although its prognosis and response to treatment are variable, the majority of patients succumb to the disease, even when given intensive treatment sometimes, including allogenic hematopoietic stem cell transplantation (allo‐HSCT). Among the many recurrent genetic abnormalities in AML, TP53 mutations are associated with the most unfavorable prognosis. Currently, no available treatment protocols are able to overcome this dismal clinical outcome [89, 90]. According to the European Leukemia Net and the National Comprehensive Cancer Center Network, even allo‐HSCT in patients with p53‐mutated AML fails to provide a long‐lasting therapeutic alternative [91, 92]. Patients with myelodysplastic syndrome (MDS) may also show TP53 mutations, and when they do, a high frequency of progression to TP53‐mutated AML also [93, 94].

Mantle cell lymphoma (MCL) is a rare subtype of B‐cell non‐Hodgkin lymphoma characterized by significant clinical and biological heterogeneity. TP53 mutations are consistently associated with a negative prognosis and exhibit a strong and independent association with early disease progression and death, particularly in patients treated with conventional intensive chemoimmunotherapy [95]. Currently, TP53 sequencing is not performed routinely in all MCL patients, and clinical trials on that subject have employed various detection methods with differing sensitivities and accuracies to identify TP53 abnormalities [96, 97]. Nevertheless, the updated guidelines from the National Comprehensive Cancer Network do recommend TP53 sequencing in all MCL patients before treatment [98].

Mutations in TP53 can also be seen in healthy individuals and give rise to clonal hematopoiesis (CH), a condition sometimes preceding leukemia, where mutated hematopoietic stem cells (HSCs) proliferate without developing overt malignancy [99, 100]. Somatic mutations in the HSCs of normal individuals drive CH and give advantages to mutant cells. The acquisition of CH‐driver mutations in genes such as DNMT3, TET2, TP53, or ASXL1 occurs with normal aging (detected in more than 10% of individuals aged ≥65 years) [101]. These clonal expansions, such as those thought to be founding events in myeloid malignancies, confer a fitness advantage to the cell.

The term therapy‐related myeloid neoplasms (t‐MNs) describes a distinct category of myeloid neoplasms (t‐AML or t‐MDS) that develop following cytotoxic or radiation therapy for solid tumors or other hematological malignancies (e.g., Hodgkin disease, non‐Hodgkin lymphoma, and multiple myeloma) [102]. There are several features that distinguish t‐MNs from de novo MNs, including a higher incidence of TP53 mutations, abnormalities of chromosomes 5 or 7, complex cytogenetics, and a reduced response to chemotherapy. In t‐MNs, TP53 mutations are not induced by the treatment itself but by the specific selection of TP53‐mutated CH cells that are resistant to DNA‐damaging therapy. These neoplasms are associated with poor prognoses [103].

Because p53 is a transcription factor, a great amount of effort has been put into analyzing the possibility of a gene‐expression signature, derived from differences between p53 mutant and wild‐type tumors, potentially able to provide a more accurate measure of the functional activity of p53. Furthermore, as an RNA‐based classification method, it may detect downstream activity changes in the TP53 signaling pathway potentially associated with specific cancer types or differentiated tumor subtypes. In breast cancer, different p53 gene signatures were shown to predict prognosis and survival in estrogen receptor‐positive breast cancer [104, 105, 106, 107]. Moreover, they have been associated with chromosomal and genomic instability and have also shown to have potential predictive value for responsiveness to immune checkpoint inhibitors [108]. Similar predictions using multiple combinations of gene signatures have also been observed in other cancer types, such as brain or bladder cancer [6, 109]. These signatures clearly show that TP53‐associated tumors are linked with alterations in gene expression pathways involving proliferation, mitosis control, or genetic integrity. However, the deployment of expression signatures in the clinic is not currently pertinent; they may add a moderate improvement in predictive accuracy when added to the current prognostic factors, but not yet a significant one.

TP53: a promising target for therapy?

The high frequency of TP53 mutations in tumors and the gene's intrinsic tumor suppressor function make it a highly promising target for tumor therapy. However, challenges arise due first to the unique structure of p53, characterized by a smooth surface lacking a suitable drug‐binding site, and second to the complexities involved in reactivating p53 functionality. Both have hindered drug development for decades [16, 17, 110]. Despite these obstacles, researchers have remained hopeful and made progress in recent years (Fig. 6). In wild‐type p53 tumors, the primary strategy is to inhibit p53 interaction with either MDM2 or MDM4 to reduce cell viability. Nutlin‐3a was the first‐discovered MDM2 antagonist capable of suppressing the growth of wild‐type p53‐expressing cancer cells of different origins in vitro and in vivo [111]. Nutlin‐3 binds to the p53 binding pocket in MDM2, thus inhibiting the p53/MDM2 interaction and leading to stabilization of p53 and induction of p53‐dependent cell cycle arrest or apoptosis. Other MDM2 antagonists with increased affinity, such as idasanutlin (RG7388), have been developed and have completed Phase I, II, and III clinical trials [110]. Unfortunately, p53 activation via MDM2 antagonists drives a strong selection for loss of p53 function and leads to resistance, as clones with p53 mutations emerge [112]. Indeed, an increase in TP53 mutation burden—likely from minor pre‐existing mutant clones—was observed weeks into clinical trials of MDM2 antagonists, correlating with resistance and limited clinical response [113, 114].

Fig. 6.

Strategies for targeting p53 in human cancer. ATO, arsenic trioxide.

The significant accumulation of mutant p53 in tumor cells makes it a prime target for the development of drugs able to partially restore the anti‐proliferative function of the protein. Indeed, that goal has been the subject of numerous studies over the past 20 years. Combinations of high‐throughput screenings of library drugs or structure‐based designs have been explored to find new molecules affecting either the thermostability, the DNA‐binding capacity, or the transcriptional activity of mutant p53.

The most clinically advanced drug targeting mutant p53 is eprenetapopt (APR‐246). This compound also completed Phase I, II, and III clinical trials, notably in patients with TP53‐mutant MDSs [115]. Eprenetapopt restores the normal tumor‐suppressor function of mutant p53 by converting it into a functional, wild‐type‐like conformation able to trigger cancer cell death through apoptosis and cell cycle arrest. It was later shown to have a broader effect unrelated to p53 by inducing oxidative stress within tumor cells, therefore enhancing its cytotoxic effects [116].

Large‐scale screenings have targeted multiple p53 variants, but structural analyses of mutant p53 have revealed some specific features, including surface cavities that could be targeted by small molecules. These latter could have the advantage of minimizing off‐target effects compared to broader p53‐reactivating compounds [117]. The cavity‐creating p53 cancer mutation p.Y220C—a hot spot variant specific for MDS and AML—has been a paradigm for developing small‐molecule drugs based on protein stabilization [118, 119]. Multiple compounds restoring wild‐type p53 protein structure and tumor suppressor function have been identified. Ongoing phase I/II clinical trials will evaluate the safety, tolerability, and anti‐tumor activities of these drugs [110].

More recently, arsenic trioxide (ATO), an established agent in treating acute promyelocytic leukemia, was shown to rescue structural p53 mutations [120]. ATO was able to extend survival in mouse models expressing mutant p53 and showing an increase in p53 activity in tumor cells [121].

Loss of p53 function can also contribute to immune evasion by influencing immune cell recruitment to the tumor, cytokine secretion in the tumor microenvironment, and inflammatory signaling pathways. This has led to interest in the clinical translation of therapies involving p53 activation to induce an immune response with or without immunotherapy [68, 69]. Pharmacological activation of p53 expression using MDM2 antagonist has been shown to abolish tumor immune invasion and promote anti‐tumor immunity in animal models [122].

Taken together, p53 alteration and its multiple consequences for the protein itself, the cell or tumor in which it is expressed and its environment, or more generally for the patient's immune response, are all potential targets for new therapies. However, there is still a long and frustrating way to go, as p53 is still considered undruggable despite being a major driver event in human cancer [16].

TP53 and germline mutations

Germline TP53 mutations have been found in patients with Li–Fraumeni syndrome (LFS), an autosomal dominant cancer predisposition syndrome characterized by the early onset of a spectrum of childhood and adulthood cancers, including adrenocortical carcinomas, soft tissue sarcomas, osteosarcomas, brain tumors, and pre‐menopausal breast cancer [123, 124, 125]. The penetrance of p53‐mediated cancer in families with this hereditary disorder is high and more pronounced in women than in men, primarily due to breast cancer. Further studies have shown that germline p53 mutations are also found in the context of hereditary breast and ovarian cancer and in childhood cancers [126, 127, 128, 129]. It is noteworthy that the landscapes of germline and somatic TP53 mutations share the same hot spot variants, suggesting a common etiology for them. Like in other TSGs, founder TP53 variants have been identified. The most prevalent one is the Brazilian variant (c.1010G>A (p.R337H)) found in 0.3% of the general southern and southeastern Brazilian population [130, 131, 132]. This variant was first detected in patients with pediatric adrenocortical tumors (ACT) and later shown to be associated with other pediatric cancers, such as choroid plexus carcinoma (CPC) [133]. Interestingly, a nonsense mutation in the TSG XAF1 close to TP53 has been shown to be associated with a more aggressive cancer phenotype when compared to the Brazilian p53 variant alone [132]. Functional analysis of p.R337H shows that it displays a very weak LOF only apparent at low pH. Mouse models carrying this variant are not prone to cancer and have the same life expectancy as wild‐type mice but were shown to be susceptible to liver tumors when they were exposed to carcinogens. The specificity of p.R337H for ACT or CPC is currently unknown.

A second founder variant (c.1000G>C; p.G334R) was observed predominantly in Ashkenazi Jews associated with low penetrance LFS, and a third, c.541C>T, (p. R181C) in Middle Eastern families with cancers characteristic of LFS [45, 46]. The weak LOF of this variant also explains the unique observation of homozygote individuals issued from consanguineous patients, each a carrier of the mutation.

Further investigation is needed to determine the potential existence of other, heretofore undescribed, genetic variants able to modify TP53 response in different tumoral or normal cellular contexts and what clinical impact they might have.

In 2011, to determine whether early tumor detection enabled quicker treatment and improved overall survival, a surveillance system based on biochemical analysis and imaging studies (Toronto protocol) was established for LFS patients [134]. Multiple subsequent studies from various groups have shown that the protocol does improve patient survival rates [135, 136, 137, 138].

TP53 polymorphisms

Multiple non‐pathogenic single nucleotide polymorphisms (SNPs) have been found scattered all along TP53 [139]. The p.P72R (rs1042522) SNP was the first to be reported and is the most frequent one found in the coding region of TP53. The very unusual features of rs1042522 have inspired extensive study of it. First, this SNP is located in the PRD of p53 and modifies protein folding [140]. Functional analyses have also shown that variant Arg72 has more potent apoptotic activity and that it may be associated with a better treatment response. Whether or not p53 variants associated with either the Pro or Arg allele display the same LOF is still debated [141]. As of today, rs1042522 detection is not used in the clinic.

Multiple, rare, benign SNPs have been identified in the coding region of p53, many of which are restricted to specific ethnicities, such as rs587780728 (p.D49H) or rs201753350 (p.V31I) in the Japanese or rs1800371 (p.P47S) and rs150607408 (p.S185N) in Africans [142]. Their clinical significance is currently unknown. A genome‐wide association study reported a novel rare variant (rs78378222) in the polyadenylation signal sequence of TP53 (AATAAA to AATACA) and showed that it impaired TP53 stability, leading to a decrease in p53 expression and hampered apoptosis [143]. This SNP, restricted to African and European populations, is strongly associated with an increased risk of cancer.

TP53 and non‐neoplastic syndromes

For several decades, studies on p53 have focused on its LOF associated with neoplastic diseases. More recent studies have shown that an increase in the functional activity of p53 could account for the clinical phenotype of many genetic diseases [144].

Inherited bone marrow failure (BMF) syndromes were initially described in different categories of disorders, each associated with germline mutations in genes defining specific molecular or functional pathways whose defect leads to cellular stress associated with p53 activation [145, 146, 147]. In telomere biology disorders, such as dyskeratosis congenita and its Revesz and Hoyeraal–Hreidarsson syndrome (RS and HHS) variants, germline mutations in genes essential for telomere biology such as DKC1, TERC, TERT, NOP10, and NHP2 lead to a defect in telomere length that can cause genetic instability [148]. Similarly, in ribosomopathies such as Diamond‐Blackfan anemia (DBA) or Shwachman–Diamond syndrome (SDS), mutations in genes coding for ribosomal proteins will lead to a dysfunctional translation of all proteins [149]. Finally, in Fanconi anemia and ERCC6L2 syndrome, mutations in DNA repair genes cause an accumulation of unrepaired DNA damage [150]. The first commonality of all these diseases is persistent and universal cell stress that exacerbates the p53/p21 DREAM pathway and leads to senescence, growth arrest, or apoptosis [151, 152, 153]. The second common trait is impaired hematopoiesis leading to anemia and BMF [149, 154]. This aligns with murine model studies suggesting a critical role for p53 in HSC self‐renewal and quiescence [155]. Therefore, precise regulation of p53 activity is likely to be important in determining the response of HSCs to cellular stresses. Insufficient p53 activation favors cell survival but puts cells at risk of losing genomic integrity. In contrast, excessive p53 activation compromises steady‐state hematopoiesis and its recovery following exogenous marrow insult by causing too many cells to be eliminated, a situation that accounts for the anemia observed in these patients.

Another common feature of these diseases is the increased risk of developing AML, MDS, and other types of cancer [149, 156]. SDS is a genetic disease associated with germline mutation in SDBS, a gene encoding a ribosomal protein. Patients with it frequently develop myeloid malignancies with unfavorable prognoses. Genetic analysis of tumor cells shows that they are always associated with somatic TP53 mutation, which explains the poor prognosis and the resistance to treatment [157]. Surprisingly, in these AML patients, TP53 mutations have also been found in non‐tumoral hematological cells several years before the development of the malignancies. These mutations likely result from CH [158]. Subsequent studies have shown that they remove the negative fitness constraint induced by the sustained activation of p53 (Fig. 5). This allows enough hematopoietic cells to survive to avoid any lethal phenotypes [159]. Unfortunately, this maladaptive somatic gene rescue will enhance the leukemic potential of these cells when the second allele of TP53 is lost. Interestingly, in SDS patients, another p53‐independent somatic genetic rescue has been observed with a somatic mutation in the ribosomal gene EIF6, which improves the translation of the deficient protein [160]. Most of these patients do not develop myeloid malignancies thanks to this non‐leukemia compensatory pathway.

Fig. 5.

Increased activation of p53 can lead to bone marrow failure syndrome and leukemia. (a) Multiple chronic cellular stresses may result in the persistent activation of p53, causing disruption of cellular homeostasis and aplasia. Although some organs can tolerate this, others, such as hematopoietic stem cells (HSC), will be deprived of specific cellular lineages. (b) In certain diseases, somatic gene rescue, such as p53 mutation, can mitigate the cellular response induced by stress but carries a risk of leading to cancer.

A similar observation has been made in the setting of patients with ERCC6L2 syndrome (associated with impaired DNA repair) who are also at high risk of developing TP53 mutated AML [161]. The compensatory p53 mutation alleviates the p53 burden associated with chronic DNA damage in hematopoietic cells.

Although p53 mutations involved in human cancer are associated with a loss of the tumor suppressive effect of TP53, there have also been observations of germline variants with increased p53 activity in patients with clinical phenotypes related to DBA, including anemia and facial dysmorphism [162, 163]. These frameshift mutations localized after the TET domain remove the NEG‐REG of p53, leading to persistent activity. It is noteworthy that variants removing this region of p53 are not observed in human tumors, suggesting counter‐selection.

In various mouse models, p53 hyperactivity has been shown to phenocopy many developmental diseases, including features of CHARGE syndrome (acronym: coloboma, heart defects, atresia choanae, growth retardation, genital abnormalities, and ear abnormalities) [164]. In most cases, patients with CHARGE syndrome carry a germline mutation in the CDH7 gene, which encodes a chromatin remodeling factor essential for proper transcriptional regulation. The CDH7 gene product was found to be a negative regulator of p53 and, therefore, accounts for the p53 hyperactivation phenotype associated with these developmental disorders. Treacher Collins syndrome (TCS) is a rare congenital craniofacial development disorder caused by mutations in TCOF1, which encodes Treacle, a nucleolar phosphoprotein that plays key roles in ribosome maturation and regulates neuroepithelial survival and neural crest cell proliferation. TCS deficiency results in nucleolar stress activation of p53, which, in turn, transcriptionally activates numerous proapoptotic response genes within the neuroepithelium, leading to the tissue‐specific death and inborn malformations observed in the pathogenesis of TCS.

Due to the central role of p53 in coordinating various cellular stresses in different cellular contexts, it is likely that other syndromes associated with p53 hyperactivity will be uncovered in the near future. Broadening settings beyond neoplastic diseases will be essential for future research.

Remaining challenges and perspectives

Despite 45 years of intense research, many grey areas remain concerning the various signaling pathways regulated by p53, the most important of which are how p53 protects multicellular organisms from cancer and how p53 LOF moves cells toward transformation. To understand this elusive tumor‐suppressive function of p53, we must move beyond the outdated “one gene, one protein, one function” model and adopt a multilayered, systems biology perspective. The complexity of p53's role in cancer stems from its ability to produce multiple isoforms through alternative splicing; thereafter, any one of those isoforms will have a distinct regulatory effect on cell fate. Additionally, PTMs—such as phosphorylation, acetylation, and ubiquitination—fine‐tune p53's function, determining whether it induces apoptosis, cell cycle arrest, or senescence. Through its pleiotropic nature, p53 is able to regulate diverse biological pathways, including metabolism, genomic stability, immune surveillance, and oxidative stress, making its tumor‐suppressive effects highly context‐dependent. Furthermore, we now know that non‐coding RNAs modulate p53's activity and that p53 itself influences the expression of tumor‐suppressive miRNAs, forming intricate feedback loops. Epigenetic modifications, such as DNA methylation and histone modification, can alter p53's transcriptional network, making it challenging to pinpoint a single mechanism of action. The tumor microenvironment further modulates p53's suppressive role, with external signals influencing whether it acts as a protector or a bystander. By considering p53 as a dynamic, context‐dependent regulator rather than a simple on/off switch, it should be possible to develop precision therapies that restore its function in a manner tailored to specific cancers. This network‐based understanding is essential for unlocking the full potential of p53 as a therapeutic target and improving cancer treatments.

TP53 mutations act within a complex network of genetic and environmental factors that shape their effects on cancer progression and treatment response. There is considerable heterogeneity among the range of p53 variants, but factoring in variant identity beyond the classical “wild‐type versus mutant” clinical status remains elusive. It is essential to keep in mind that the impact of TP53 mutations on cancer is shaped by both the tumor's and the individual's genetic background. In the tumor, factors such as co‐occurring mutations, epigenetic modifications, and DNA repair deficiencies—just to name a few—influence how TP53 mutations drive malignancy. Additionally, p53 interacts with key signaling pathways (e.g., PI3K/AKT, Wnt) and is affected by aspects of the tumor microenvironment, including hypoxia and inflammation. In the individual, germline TP53 mutations (e.g., in LFS) increase cancer risk, whereas some polymorphisms, either in TP53 or other genes, modify cancer susceptibility and treatment response. Ethnic differences in TP53 variants, immune system variability, and metabolic factors can further influence tumor behavior. In hormone‐driven cancers, estrogen or androgen signaling affects TP53’s role. Aging‐related changes and inherited genetic syndromes (e.g., ataxia‐telangiectasia) can also alter TP53 function. Moreover, the consequences of TP53 mutations can be affected by environmental factors (e.g., smoking, diet, microbiome composition) and individual differences in DNA repair efficiency or drug metabolism. Even epigenetic inheritance can play a role, as parental exposures might modify TP53‐related pathways across generations. Thus, each p53 variant acts within a complex network of genetic and environmental factors that shape its effect on cancer progression and treatment response.

Author contributions

Thierry Soussi and Panagiotis Baliakas contributed equally to the conception and design, drafting and critical revision of the manuscript.

Conflict of interest statement

The authors declare no conflicts of interest.

Baliakas P, Soussi T. The TP53 tumor suppressor gene: From molecular biology to clinical investigations. J Intern Med. 2025;298:78–96.