TP53, TP63, and TP73 genes comprise the p53 family. Each gene produces protein isoforms through multiple mechanisms including extensive alternative mRNA splicing. Accumulating evidence shows that these isoforms play a critical role in the regulation of many biological processes in normal cells. Their abnormal expression contributes to tumorigenesis and has a profound effect on tumor response to curative therapy. This paper is an overview of isoform diversity in the p53 family and its role in cancer.

1. Introduction

Alternative splicing allows a single gene to express multiple protein variants. It is estimated that 92–95% of human multiexon genes undergo alternative splicing [1, 2]. Abnormal alterations of splicing may interfere with normal cellular homeostasis and lead to cancer development [35].

The p53 protein family is comprised of three transcription factors: p53, p63, and p73. Phylogenetic analysis revealed that this family originated from a p63/73-like ancestral gene early in metazoan evolution [6, 7]. Maintenance of genetic stability of germ cells seems to be its ancestral function [8]. The p53 family regulates many vital biological processes, including cell differentiation, proliferation, and cell death/apoptosis [9, 10]. Dysregulation of the p53 family plays a critical role in tumorigenesis and significantly affects tumor response to therapy. This review summarizes current data on the regulation of p53, p63, and p73 isoforms and their roles in cancer.

2. Structure and Function

p53, p63, and p73 genes are located on chromosomes 17p13.1, 3q27-29, and 1p36.2-3, respectively. These genes encode proteins with similar domain structures and significant amino acid sequence homology in the transactivation, DNA-binding and oligomerization domains (Figure 1). The highest amino acid identity is in the DNA-binding domain (~60%). Evolutionally, this domain is the most conserved, suggesting that regulation of transcription plays a pivotal role in an array of functions attributed to the p53 family. Less similarity is found in the oligomerization and transactivation domains (~30%).

The founding member of the p53 family, the p53 protein, had been discovered more than three decades ago [12, 13]. For a long time, it had been assumed that p53 is expressed as a single polypeptide. However, when it had been found that the p63 and p73 genes encoded a large variety of diverse transcripts, the p53 gene transcription was revisited. Now we know that p53 forms multiple variants.

Transcriptions of p53, p63, and p73 genes are regulated by similar mechanisms. It is controlled by two promoters: P1 and P2, where P2 is an alternative intragenic promoter (Figure 1). One study in silico provided evidence for the existence of a third putative promoter in the first intron of human TP73 gene [14]. Therefore, it would not be surprising if additional gene promoters will be found in the future. An extensive alternative splicing adds further diversity to the promoters’ products. The produced transcripts and proteins can be generally categorized into two main groups, termed TA and ΔN [15, 16]. TA variants contain the N-terminal transactivation domain while ΔN isoforms lack the entire (or part of) domain. It was initially thought that ΔN isoforms are only generated by the P2 promoter whereas the P1 promoter regulates TA isoforms. Further analysis of alternative mRNA splicing revealed that some transcriptionally deficient isoforms are products of the P1 promoter. For example, the P1 promoter of the TP73 gene regulates TAp73 isoforms and isoforms, which lack the TA domain: ΔEx2p73, ΔEx2/3p73, and ΔN′p73. The latter isoforms are missing either exon 2 (ΔEx2p73) or both exon 2 and 3 (ΔEx2/3p73) or contain an additional exon 3′ (ΔN′p73) [17, 18]. Other ΔNp73 transcripts are products of the P2 promoter. Similar to p73, the P1 promoter of the p53 gene produces transcriptionally active isoforms [5]. The alternative splicing is responsible for transcriptionally deficient isoforms of Δ40p53, which missing the first 40 amino acids at the N-terminus [5, 19, 20]. Additional p53 transcriptionally deficient isoforms (Δ133p53 and Δ160p53) are regulated by the P2 promoter located in intron 4 of the p53 gene [5, 21].

Additional diversity of p53, p63, and p73 transcripts is generated by alternative splicing at the 3′ end of the transcripts (Figure 1). These splice variants are traditionally named with letters of the Greek alphabet. Initially, three such splice variants have been described for p63 and p53 (α, β, γ), and nine for p73 (α, β, γ, δ, ε, θ, ζ, η, and η1) [2225]. Later, additional p63 splice variants (δ, ε) and p53 (δ, ε, ζ, ΔE6) were reported [2628]. However, it should be noted that a majority of p53, p63, and p73 studies focus on a few isoforms, primarily α, β, and γ. Little is known about the functions of other isoforms. The combination of alternative splicing at the 5′ and 3′ ends, alternative initiation of translation and alternative promoter usage can significantly increase protein diversity. For example, N-terminal variants (p53, Δ40p53, Δ133p53, and Δ160p53) can be produced in α, β, and γ “flavors” [20, 21]. Theoretically, the p53 gene can produce at least 20 isoforms, p63 at least 10, and p73 more than 40, though not all have been experimentally confirmed.

p53, TAp63, and TAp73 share significant functional resemblance. They can induce cell cycle arrest, apoptosis, or cellular senescence. This similarity can be explained, at least in part, by transactivation of the same transcriptional targets. Genome-wide analyses found an overlap of the transcription profiles of p53, TAp73, and TAp63, though unique targets were identified as well. Analyses using chromatin immunoprecipitation, reporter, and gel-shift assays found that TAp73 and TAp63 interact with p53-responsive elements.

The transactivation and apoptotic potential of p53, TAp73, and TAp63 vary greatly depending on the isoform. TAp63γ and TAp73β are similar to that of p53α [29]. Other isoforms are considered less active on the p53 target gene promoters [9, 23, 30]. Some isoforms are characterized by a variation in domain structure. TAp73α and TAp63α have an additional domain at the COOH-terminus that is not found in p53. This domain, termed SAM or Sterile Alpha Motif, is responsible for protein-protein interactions and is found in a diverse range of proteins that are involved in developmental regulation. It is also implicated in transcriptional repression [31]. Beta and gamma isoforms of p53 are missing most of the oligomerization domain that results in decreased transcriptional activity [5, 32, 33].

ΔN isoforms function as dominant-negative inhibitors of TA counterparts (Figure 2). Promoter competition and heterocomplex formation have been suggested to explain this phenomenon [17, 34, 35]. In the promoter competition mechanism, the suggestion is that ΔN competes off TA isoforms from their target gene promoters, thus preventing efficient transcription. In the heterocomplex formation mechanism, ΔN isoforms would inhibit TA by forming hetero-oligomeric complexes.

ΔN isoforms of p53 and p73 are regulated by a negative feedback loop mechanism. Analogous mechanism was not described for p63 despite its significant similarity to p73. In a nutshell, TA isoforms are able to induce transcription of ΔN isoforms by activating P2 promoters. The induced ΔN isoforms, in turn, inhibit TA isoforms. A good example of these interactions is an induction of Δ133p53 by p53 [5, 3638]. Similarly, TAp73 and p53 are important regulators of transcriptions of ΔNp73 [39]. It appears that the balance between ΔN and TA isoforms is finely tuned to regulate the activities of TA isoforms. The net effect of these interactions in a given context appears to be dependent on the TA/ΔN expression ratio. Deregulation of this mechanism may lead to tumor development [4042]. However, it has become clear that the role of ΔN isoforms is multifaceted. The dominant negative concept cannot explain the complexity of all the interactions attributed to ΔN isoforms. Several studies reported that ΔN isoforms can retain transcription activity through additional transactivation domains.

3. Role of p53 Isoforms in Cancer

Although many aspects of p53 biology have been thoroughly investigated, the role and regulation of p53 isoforms remain not well understood.

Recent studies suggested that Δ133p53 isoform may play an oncogenic role. Mice overexpressing the Δ122p53 isoform (murine homolog of human Δ133p53) show reduced apoptosis, increased cell proliferation and develop a wide-spectrum of aggressive tumors including lymphoma, osteosarcoma, and other malignant and benign tumors [43]. Another phenotypic characteristic of these mice is elevated cytokine levels in the blood and widespread inflammation in many organs. Interestingly, transgenic expression of another p53 isoform, Δ40p53, does not lead to tumor formation in mice, but is associated with a short life span, cognitive decline, and overt diabetes, suggesting a significant difference between these isoforms [4446].

Several studies reported an elevated expression of Δ133p53 in tumors (Table 1). In breast tumors, 24 of 30 cases showed an increased expression of Δ133p53, but low or undetectable levels in normal breast tissue [5]. An increase of Δ133p53α mRNA was also found in renal cell carcinoma [47]. In colon tumors, progression from colon adenoma to carcinoma is accompanied by an increase of Δ133p53 mRNA. This study suggested that Δ133p53 helps to escape from the senescence barrier during colon tumor progression [48]. Interestingly, the Δ133p53 expression level is associated with the mutation status of p53; colon tumors expressing wildtype p53 had higher levels of Δ133p53 than p53 mutant tumors [48]. In addition to Δ133p53, an increased expression of Δ40p53 was also reported in human melanoma cell lines and primary melanoma isolates [33]. However, not all tumors overexpress Δ133p53. Analysis of squamous carcinoma of the head and neck did not reveal any significant changes in the Δ133p53 levels, suggesting that this isoform may only play a tumor-promoting role in a subset of tissues [49].

Alterations of p53β and p53γ isoforms were also reported in different types of cancers (Table 1). An increased expression of p53β was found in renal cell carcinoma and in most melanoma cell lines. In renal cell carcinoma, p53β expression was associated with tumor progression [47]. p53β was also found to correlate with worse recurrence-free survival in ovarian cancer patients with functionally active p53 [28]. Decreased p53β and p53γ mRNA levels were reported in breast cancer [5]. In breast tumors, p53β is associated with the expression of estrogen receptor but not with disease outcome [50]. Breast cancer patients expressing both mutant p53 and p53γ have lower cancer recurrence and favorable prognosis [51]. Currently, specific functions of p53β and p53γ remain unclear. A significant hurdle to the studies of p53 isoforms in tumors is the lack of isoform-specific antibodies. The generation of new antibodies, animal models, and additional tumor studies may help to better understand the role of p53 isoforms in tumorigenesis.

4. Role of p73 Isoforms in Cancer

The role of p73 in tumorigenesis is still a matter of debate. In contrast to p53, p73 is rarely mutated and frequently overexpressed in human tumors [23, 5256]. An initial study of p73-deficient mice found a number of developmental defects and no spontaneous tumors [57]. Follow-up studies have revealed spontaneous tumorigenesis, although the late onset of tumors and smaller tumor sizes compared to p53-deficient animals were reported. The basis for these conflicting results in cancer susceptibility remains obscure but might be related to the animal genetic background and housing conditions. Mice with isoform-specific knockouts of p73 have also been generated; phenotypes of these animals generally reflect previously reported differences between p73 isoforms. TAp73 null mice are tumor prone while ΔNp73 knockouts have increased sensitivity to DNA-damaging agents and elevated p53-dependent apoptosis [58, 59].

Several studies have found that N-terminally truncated isoforms of p73 play an oncogenic role and are linked to cancer development (Table 1). Targeted transgenic overexpression of human ΔEx2/3p73 in the mouse liver resulted in the development of hepatocellular carcinoma [60]. The N-terminally truncated isoforms are upregulated in many human cancers including liver, ovarian, breast, vulvar cancers, and melanoma [23, 6168]. Overexpression of ΔEx2p73 and ΔEx2/3p73 was found to be associated with metastases in melanoma [68].

ΔNp73, which is produced by the P2 promoter, has also been found to behave as an oncogene. ΔNp73 facilitates immortalization of primary mouse embryonic fibroblasts and cooperates with oncogenic Ras in their transformation. These transformed cells produce tumors following a subcutaneous injection into nude mice [121, 122]. ΔNp73 also inhibits differentiation of myoblasts and protects them against apoptosis [123]. Studies by others and us found that ΔNp73 is upregulated in a number of tumors and is associated with metastases, chemotherapeutic failure, and poorer patient prognosis [62, 74, 96, 124130].

An important question is what causes deregulation of p73 isoforms in tumors? One of the mechanisms is tumor-specific alternative mRNA splicing. It has been demonstrated that the alternative splicing causes incorporation of a new exon 3’ into TAp73 transcripts resulting in a translational switch from TAp73 to ∆Np73 isoform [18, 61]. An interesting observation was also made in hepatocellular carcinoma where an aberrant switch from TAp73 to ΔEx2p73 was found to be mediated by the activation EGFR by amphiregulin. This leads to activation of JNK1 kinase, suppression of splicing factor Slu7, and alternative splicing of p73 transcripts [65]. Activated Ras has also been shown to decrease TAp73 levels and increase ΔNp73 expression during cellular transformation [131]. Abnormal regulation of the P2 promoter has also been reported. We found that transcriptional repressor HIC1 (Hypermethylated In Cancer 1) can suppress expression of ΔNp73 by inhibiting the P2 promoter in normal cells. Loss of HIC1 in esophagus and gastric cancer cells leads to up-regulation of ΔNp73 [96]. In a subset of tumors, abnormal epigenetic changes cause deregulation of p73 isoforms [132134]. Hypomethylation of the P2 promoter was found in more than half of non-small lung cancers [76].

An increased expression of TAp73 isoforms was also found in tumors, although its role remains unclear (Table 1). Several studies suggested that in specific circumstances TAp73 might play a tumor-promoting role [30, 135]. Interestingly, some tumors tend to increase a variety of p73 splice isoforms (Figure 3). In the normal colon and breast, p73α and p73β isoforms are predominant whereas other spliced variants (γ, δ, ϕ, and ε) are primarily detected in colon and breast cancers [15, 23]. This phenomenon was also observed in acute myeloid leukemia. Moreover, the p73ε isoform was only expressed in leukemic cells and completely absent in mature myeloid cells [136]. It is currently unclear what role these changes play in tumorigenesis.

5. Role of p63 Isoforms in Cancer

Similar to p73, mutations in the p63 gene are rare in human cancers [90, 137, 138]. Several studies reported that ΔNp63 has oncogenic properties. Ectopic overexpression of ΔNp63 in Rat-1A cells promotes colony formation in soft agar. When xenografted into immunocompromised mice, these cells formed tumors [139]. ΔNp63α inhibits oncogene-induced cellular senescence and cooperates with Ras to promote tumor-initiating stem-like proliferation [140]. Analysis of p63-deficient mice led to conflicting results with regard to the p63 role in tumorigenesis. p63-/- null mice showed striking developmental defects demonstrating a critical role of p63 in epithelial development [141, 142]. p63+/- heterozygous mice were shown to be susceptible to tumor development [143]. However, other mouse models were not consistent with this observation. Conflicting phenotypes of TAp63 and ΔNp63 transgenic mice have also been reported [144, 145].

ΔNp63 is a predominant isoform expressed in most epithelial cells. Overexpression of ΔNp63 is found in cancers of nasopharyngeal, head and neck, urinary tract, lung, and ovarian tumors and correlated with poor outcome [78, 146149]. In metastases, ΔNp63 expression was found to be reduced or lost [91, 101]. Microarray analyses revealed the up-regulation of genes associated with tumor invasion and metastasis in p63-deficient cells [150]. It was also reported that p63 suppresses the TGFβ-dependent cell migration, invasion, and metastasis [151]. This suggests that ΔNp63 plays a dual role by promoting tumor development but suppressing metastases [151, 152]. Expression of ΔNp63 was found to be associated with an increased chemoresistance in a subset of breast and head and neck tumors [153, 154].

TAp63 isoforms induce cellular senescence and inhibit cell proliferation [155157]. TAp63 deficiency increases proliferation and enhances Ras-mediated oncogenesis [155]. Decreased TAp63 expression is associated with metastasis in bladder and breast cancers as well as poor outcome [42, 90, 158]. TAp63 impedes the metastatic potential of epithelial tumors by controlling the expression of a crucial set of metastasis suppressor genes [151, 159].

Clearly, additional studies are needed to understand the complex regulation of p63 isoforms.

6. Interplay of p53/p63/p73 Isoforms in Human Cancers

Interactions between members of the p53 family and their isoforms have a profound effect on tumorigenesis and anticancer drug response. Perhaps, the most studied are interactions between ΔN and TA isoforms. Inhibition of TAp73 by ΔNp63 has been shown to negatively affect the response to platinum-based chemotherapy in head and neck squamous cell carcinomas and a subset of breast tumors [153, 154]. In carcinomas of ovary and childhood acute lymphoblastic leukemia, increased expression of dominant-negative p73 isoforms correlates with resistance to conventional chemotherapy [129, 130]. Moreover, ΔNp73 is primarily expressed in ovarian tumors, which express wildtype p53 [64]. However, crosstalk between the p53 family members is not limited to dominant-negative interactions. Accumulating evidence suggests that the p53 family interacts on multiple levels comprising protein-protein interactions between multiple p53, p63, and p73 isoforms, shared regulation of target genes as well as TP53 and TP73 gene promoters [160163]. In addition, mutant p53 can affect activities of TAp73 and TAp63. It has been shown that certain tumor-derived p53 mutants (R175H, R248W, Y220C, R249S, R283H, and D281G) can physically associate and inhibit activation of TAp73 and/or TAp63 [164166].

Current analyses suggest that the function of a particular isoform needs to be investigated in the context of expression of other isoforms. For example, ΔNp73β inhibits p53-dependent apoptosis in primary sympathetic neurons [167], but when overexpressed in cancer cells, ΔNp73β induces cell cycle arrest and apoptosis [168].

An interesting observation has been made in mouse embryonic fibroblasts, where the combined loss of p73 and p63 results in the failure of p53 to induce apoptosis in response to DNA damage [169]. More recent studies have reported that the p53 family members can simultaneously co-occupy the promoters of p53 target genes and regulate their transcription [15, 170, 171]. Notably, the integral activity of the entire p53 family, as measured by reporter analysis, is a better predictor of chemotherapeutic drug response than p53 status alone [15].

7. Conclusion

The p53 family plays a pivotal role in the control of many critical cellular functions. In recent years, it has been revealed that all members of the p53 family are expressed as a diverse variety of isoforms. We only just started to uncover the mechanisms that regulate this diversity. A number of studies also provided the first glimpses of their functional significance. Clearly, isoforms add a new level of functional regulation to many critical biological processes including cell death, proliferation, cell cycle control, and tumorigenesis. Depending on the isoform expressed, the role of a gene can dramatically change from a tumor suppressor to an oncogene. It is also clear that p53, p73, and p63 isoforms tightly interact. A better understanding of this interacting network and its regulation holds the key to future therapeutic benefits.


The authors thank Dr. El-Rifai for the valuable discussions. This paper was supported by the National Cancer Institute Grants NIH CA138833 and NIH CA108956.