How the TP53 Family Proteins TP63 and TP73 Contribute to Tumorigenesis: Regulators and Effectors

TP63: Structural and Functional Features

TP63 is the oldest evolutionary conserved homologue of the TP53 family that shares some biological functions with TP53. However, TP63 also exhibits a unique role in the regulation of epithelial morphogenesis and adult epithelial stem cell maintenance and differentiation [Mills et al., 1999; Yang et al., 1999; Candi et al., 2007b; Paris et al., 2012]. TP63 has been shown to be involved in all aspects of tumorigenesis and cancer progression including inhibition of cancer metastasis [Melino et al., 2002; Melino, 2011].

The 3′ end of both TATP63 and ΔNTP63 may be alternatively spliced to yield the C-terminal protein variants α, β, γ, and δ, while the ε-isoform is generated from transcriptional termination in exon 10 [Yang et al., 1998; Mangiulli et al., 2009] (Fig. 1B). The TP63α and TP63β isoforms contain alternative TADs (TA2) encoded in exons 11 and 12, which may be responsible for mediating ΔNTP63 transactivation activity [Ghioni et al., 2002]. Indeed, while ΔNTP63 isoforms were initially described as simple dominant-negative proteins toward the transcriptionally active family members, they can act themselves as transcription factors. The TP63α isoforms display two additional C-terminal structural domains, a Sterile alpha motif (SAM) domain and transactivation inhibitory domain (TID). The SAM domain mediates protein–protein interactions that are essential for tetramerization, and therefore for the formation and stabilization of the active molecule [Ou et al., 2007]. The TID is thought to interact with the TAD of another molecule, hence forming a closed and inactive dimeric conformation, which inhibits the transactivation properties of TP63 isoforms.

TATP63 is mainly involved in protection against cancer metastasis [Su et al., 2010; Melino, 2011] and in preserving the genomic integrity of female germ cells [Suh et al., 2006], whereas ΔNTP63, which is predominantly expressed in stratified epithelial and glandular tissues, is essential for maintaining the proliferative potential of the stem/progenitor cells in these compartments [Senoo et al., 2007; Pignon et al., 2013]. This function is well-supported by the observations that genetic ablation of ΔNTP63 in mice [Romano et al., 2012], as well as mutations in patients [Vanbokhoven et al., 2011], are associated with a plethora of severe stratified epithelial defects, which can be ascribed to a failure in maintaining the stem progenitor cell populations [Masse et al., 2012]. The TP63-deficient animals display complete lack of all stratified epithelia such as epidermis, oral and bladder epithelium, as well as glandular structures, including thymus, prostate, and breast.

The TP63 gene is rarely mutated in human cancers. This observation does not support a classical tumor-suppressor function for TP63 [Melino et al., 2002; Melino, 2011]. However, TP63 has been found overexpressed in basal cell and squamous cell carcinomas of the head and neck, generally due to chromosomal amplification [Candi et al., 2007a]. Elevated levels of TP63 are also found in poorly differentiated cervical and basal-like breast tumors, in prostate and bladder carcinomas, thymomas. Overexpression of TP63 has also been recently reported in both primary and metastatic melanoma clinical samples, where it significantly correlated with a worse prognosis [Matin et al., 2013]. ΔN proteins are the most widely expressed isoforms in tumors. Genetic studies in mice have provided conflicting results on the role of TP63 in tumorigenesis. According to Flores et al. (2005), TP63 haploinsufficiency predisposes animals to develop spontaneously a broad spectrum of tumors, though with late onset. A higher tumor penetrance and increased incidence of metastasis was observed when TP63 loss was set in a TP53 heterozygous background. Conversely, Keyes et al. (2006) reported that TP63 heterozygous mice neither develop spontaneous tumors, nor are susceptible to chemically induced carcinogenesis. Specific genetic ablation of TATP63 has unveiled a role for this isoform in the regulation of cellular senescence, which acts as a barrier against tumor progression, and confirmed the observations by Flores et al. (2005). Indeed, loss of TATP63 enhances frequency of sarcoma occurrence in mice lacking TP53 [Guo et al., 2009]. Additional TP63 isoform-specific models are required to definitively establish the contribution of each TP63 variant to tumor development, progression, and drug sensitivity.

Regulation of tissue integrity and cell adhesion by TP63 may affect tumor cell migration invasion and metastasis. Loss of TP63 has been indeed associated with increased potential to metastasize [Su et al., 2010]. In addition, TGF-β-induced inactivation of TP63 achieved through the formation of mutant TP53–TP63–Smad complexes contributes to tumor invasiveness and metastasis [Adorno et al., 2009].

TP73 Gene, Protein Structure, and Function

TP73 plays a role in several biological process such as cell death [Gong et al., 1999; Melino et al., 2004; Al-Bahlani et al., 2011], differentiation [De Laurenzi et al., 2000; Billon et al., 2004; Agostini et al., 2011a], neuronal stem cell maintenance [Agostini et al., 2010; Talos et al., 2010], aging and metabolism [Rufini et al., 2012] through the regulation of gene [Melino et al., 2004; Graupner et al., 2011] and microRNA (miR) expression [Agostini et al., 2011a; Agostini et al., 2011b] (Fig. 2).

The human TP73 gene [Kaghad et al., 1997] resides in the distal short arm of chromosome 1. This region, referred as 1p36, is frequently deleted in a broad range of tumors such as neuroblastoma, breast cancer, cervical cancer, pancreatic cancer, thyroid cancer, hepatocellular cancer, colorectal cancer, lung cancer, glioma, melanoma, rhabdomyosarcoma, acute myeloid leukemia, chronic myeloid leukemia, and non-Hodgkin lymphoma.

A schematic representation of the TP73 gene structure is shown in Figure 1C. The TP73 gene, potentially transcribes 35 mRNA variants, which can be translated theoretically in 29 different TP73 proteins isoforms [Murray-Zmijewski et al., 2006]. Overall, 14 different TP73 proteins have been described so far. As mentioned above, N-terminal truncated TP73 isoforms have been also identified. These isoforms, collectively called ΔTATP73, originate from alternative splicing events targeting the 5′ of the transcript, generated from the P1 promoter. The main products include ΔEx2TP73 (lacking exon 2), ΔEx2/3TP73 (lacking exons 2 and 3), and ΔΝ′TP73 [Melino et al., 2002]. The alternative splicing that occurs at the 3′ end of the transcript generates at least seven different mRNA variants (α, β, γ, ζ, δ, ε, and η). However, whether these isoforms are translated into proteins is still debatable. Similarly to TP63, the longest C-terminal α-isoform contains a SAM domain. Recently, the oligomerization state of TATP73α has been characterized [Luh et al., 2013]. Although TATP73α and TATP63α share the same domain organization, TATP73α forms constitutively open active homotetramers, whereas TP63 forms closed inactive homodimers. Indeed, TATP73α expressed in mammalian cells forms a higher-order species in solution that is similar in size to the open activated tetramer of TATP63α. These findings could, at least in part, explain why TP73 and TP53 are already active but expressed at low levels, whereas TATP63 is constitutively expressed at high levels such as in oocytes [Suh et al., 2006]. Moreover, the first crystal structure of the DBD of TP73 has been recently solved [Ethayathulla et al., 2012]. TP73-dependent transcription is influenced by several factors. In particular, the spacer separating the two decamer half-site regulates the activation of apoptotic or nonapoptotic pathways. This distance between the two half-site responsive elements is more important for TP73 than for TP53 [Jordan et al., 2008]. Indeed, the effect on the transactivation activity of TP73β is drastically affected by the responsive elements spacer length and also by the sequence. This suggests that the TP73 transactivation mechanism is dependent on structural changes that take place in the oligomerization interface of the tetramer upon binding to different response elements. These new findings might explain why TP53 family members, although having a highly conserved DBD and similar responsive element specificity, display different patterns of gene expression.

Biologically, the two most representative isoforms, TATP73 and ΔNTP73, show distinct functions. Similarly to TP53, TATP73 induces cell cycle arrest and apoptosis and regulates genomic stability, thus contributing to suppress tumorigenesis [Tomasini et al., 2008; Rufini et al., 2011]. The tumor-suppression function was also confirmed by in vivo studies [Tomasini et al., 2008]. Indeed, TATP73-selective knockout mice show an increased susceptibility to spontaneous and induced carcinogenesis, with more than 70% of these animals developing tumors, which mainly comprise lung adenocarcinoma. In addition, this mouse model manifests infertility and hippocampal digenesis, indicating a role of TATP73 isoform in the regulation of reproduction and neuronal development [Yang et al., 2000; Killick et al., 2011; Niklison-Chirou et al., 2013]. On the contrary, ΔNTP73 displays antiapoptotic, oncogenic activities, which are achieved through the formation of heterocomplex with TATP73, or via competition for promoter binding with both TATP73 and TP53 [Stiewe et al., 2002; Zaika et al., 2002]. The dominant-negative effect of ΔNTP73 toward the transcriptionally active TP53 proteins results in the promotion of cell survival. In support of this, tumor development has not been observed in ΔNTP73-selective knockout mice and E1A/Ras-transformed ΔΝTP73 knockout mouse fibroblasts fail to form tumors when transplanted into immunocompromised recipient mice. Moreover, ΔNTP73-deficient mice show enlarged ventricles and neuronal loss, confirming its prosurvival role in differentiated mature neurons [Pozniak et al., 2000; Wilhelm et al., 2010].

Although the chromosomal region where TP73 is located is frequently deleted in several human cancers, the picture that emerges from genetic and expression studies is more complex. Three main points can be summarized as follows: (1) TP73 is rarely mutated in cancer. Indeed, genetic data show that it is mutated in only 0.6% of primary tumors; (2) the expression of TATP73 isoforms is often deregulated in human cancers. The overall picture indicates a different pattern of isoforms expression between normal and malignant tissues. In particular, it was observed a shift toward low-molecular-weight TATP73 C-terminal isoforms expression in malignant tissue; (3) ΔNTP73 isoforms are found frequently upregulated in several primary tumors such as neuroblastoma, and ovary, breast, lung, and colon carcinomas.

Two relevant questions in the field of the TP53-like proteins are: (1) how TP63 and TP73 are controlled to achieve an isoform-dependent regulation of their biological functions in response to cellular stress (e.g., DNA-damaging agents, nutrient deprivation) and (2) which are the transcriptional programs and signaling pathways, downstream of TP63/TP73 activation, that link them to tumorigenesis? Compelling evidence is accumulating on both sides, creating a complex picture, in which regulators and targets cross-talk each other to finely tuning TP63/TP73 transcriptional function.

Posttranscriptional Regulation

Posttranscriptional regulation of the TP53 family proteins by miRs is a relevant mechanism underlying their functional activation. A broader picture is now emerging in which TP63 expression is regulated by specific miRs. miR-203 was identified as the first miR able to regulate TP63 in normal epithelia [Aberdam et al., 2008; Lena et al., 2008; Viticchie et al., 2012]. A recent study has identified miR-574–3p and miR-720, as two important regulators of TP63 through the inhibition of protein translation [Chikh et al., 2011]. Interestingly, the expression of miR-574–3p and miR-720 is negatively controlled by (inhibitor of apoptosis-stimulating protein of p53) IASPP, an inhibitory member of the ASPP (apoptosis-stimulating protein of TP53) protein family. Remarkably, TP63 transcriptionally activates the expression of IASPP, hence establishing a positive autoregulatory feedback loop, which sustains TP63 expression, and ultimately regulates the proliferative potential of epidermal basal progenitors. Under pathological conditions affecting skin, such as the chronic familial pemphigus of Hailey–Hailey, the expression of TP63 is suppressed by miR-125b in lesion skin-derived keratinocytes. miR-125b is thought to induce differentiation defects in Hailey–Hailey keratinocyte partly through the inhibition of TP63 [Manca et al., 2011].

Scheel et al. (2009) described the miR-302 cluster, and in particular miR-302, as novel antagonists of TP63 expression in cells of the germline, with possible implication in cancer. In a model system of murine diploid myeloid cells, miR-92, part of the mir-17–92 cluster, increases cell proliferation by negative regulation of TP63 [Manni et al., 2009]. Additionally, TATP63 is an important target for miR-21, a cancer-related miR with oncogenic potential, which is aberrantly expressed in glioblastoma. Remarkably, the cell cycle arrest phenotype elicited by miR-21 downregulation in glioblastoma cells, is, at least partially, dependent on TATP63 [Papagiannakopoulos et al., 2008]. Finally, Antonini et al. (2010) reported that TP63 directly binds to TP53-consensus sites in both miR-34a and miR-34c regulatory regions, and promotes cell cycle progression by directly repressing miR-34a and miR-34c. It is not yet clear whether this mechanism that sustains epithelial proliferation could also have a role in cancer formation.

Regulation of Protein Stability

E3 ubiquitin ligases (E3) are key components of the ubiquitin proteasome system, which plays a critical role in cellular protein homeostasis. Although TP63, TP73, and TP53 are structurally related, the molecular mechanisms regulating their stabilization are substantially different. Similarly to TP53, the protein steady-state levels of the majority of the other family members are kept low under unstressed condition. There is an increasing number of E3s targeting TP63/TP73 for protein ubiquitylation and subsequent degradation, which, in turn, is negatively linked to their transcriptional activation. The existence of multiple E3s indicate the importance of maintaining tightly regulated the levels of TP63/TP73 in order to finely modulate their antitumor activities [Conforti et al., 2012].

While MDM2, the main regulator of TP53 stability, does not target TP73 and TP63 for protein ubiquitylation, there is evidence that it is capable of binding the TADs of both transcription factors [Ongkeko et al., 1999; Zdzalik et al., 2010]. Interaction of MDM2 with TP73 results in suppression of its transactivating properties [Zeng et al., 1999], whereas contradictory outcomes of the MDM2/TP63 association have been reported [Kadakia et al., 2001; Calabro et al., 2002; Galli et al., 2010]. A primary regulator of TP63/TP73 protein levels is the HECT (homologous to the E6-AP carboxyl terminus)-type E3, ITCH [Melino et al., 2008], which binds them via interaction with their C-terminal proline-rich (PY) motif, and promotes their proteasomal degradation [Rossi et al., 2005; Rossi et al., 2006]. In contrast, ICTH does not interact and modify TP53. In response to DNA damage, ITCH inhibition results in TP63/TP73 stabilization and activation [Rossi et al., 2005; Rossi et al., 2006]. ITCH downregulation upon stress can be achieved through reduced transcriptional activation of the Itch promoter by RUNX [Levy et al., 2008b]. ITCH levels are also reduced by posttranscriptional mechanisms. Indeed, miR106b has been shown to repress ITCH translation and cause TP73 accumulation and functional activation [Sampath et al., 2009]. In addition, ITCH inactivation can be accomplished through competition mechanisms. As an example, the NEDD4-binding partner 1 (N4BP1) competes with TP73 for binding to the WW2 domain of ITCH, thus reducing its ability to recruit and modify the transcription factor [Oberst et al., 2007]. Similarly, Yes-associated protein 1 (YAP1) stabilizes TP73 by competing with ITCH for interaction with overlapping binding sites [Levy et al., 2007]. The association of YAP1 and TP73 is indeed mediated by the WW domain of YAP1 and the PY motif of TP73. These mechanisms allow TP73 to escape ITCH-mediated ubiquitylation and subsequent degradation.

Another member of the HECT family of E3s, WW domain-containing protein 1 (WWP1) has been shown to target TP63 for protein ubiquitylation [Li et al., 2008; Peschiaroli et al., 2010]. Similarly to ITCH, binding of WWP1 to TP63 is mediated by the C-terminal PY motif of TP63 [Li et al., 2008]. Interestingly, the outcome of TP63 modification by WWP1 may vary depending on the targeted isoform, cell type, and nature of ubiquitin linkages. Modification of ΔNTP63 through conjugation of Lys63-linked polyubiquitin chains promotes its transcriptional activation, but does not affect protein stability [Peschiaroli et al., 2010]. Conversely, Li et al. (2008) have reported that WWP1-mediated ubiquitylation of TP63 controls protein stability of either TATP63 or ΔNTP63 variants in a cell line-dependent fashion. Accordingly, TP63 degradation may confer increased or reduced sensitivity to drug-induced apoptosis [Li et al., 2008]. Involvement of WWP1 in TP73 regulation has not been described so far, whereas modification of TP53 by WWP1 seems to promote its nuclear export and cytoplasmic sequestration, which, in turn, results in its transcriptional inactivation [Laine and Ronai, 2007].

TP63/TP73 are also target of RING-finger domain-type E3s. TP73 stability is controlled by an SCF E3 complex containing the F-box FBXO45 [Peschiaroli et al., 2009], the human ortholog of C. elegans FSN-1, which regulates the TP53-like protein CEP-1. TP73 is the only family member, which is modified by FBXO45 in mammalian cells. The SCF complex SCF(βTrCP1) binds the TP63γ isoforms, with a higher affinity for the TATP63γ variant [Gallegos et al., 2008]. Interaction of SCF(βTrCP1) with TATP63γ leads to protein accumulation and increased activation of TP63 target genes.

Similarly to the TP53/MDM2 regulatory loop, there are many examples of mutual regulation between an E3 and the TP63/73 proteins. As an example, the E3 tripartite motif protein 32 (TRIM32) is a recently described direct transcriptional target of TP73 [Gonzalez-Cano et al., 2013], which is differentially regulated by distinct TP73 isoforms, with TATP73 inducing, whereas ΔNTP73 repressing its expression. TRIM32 then promotes TP73 ubiquitylation and degradation, thus reducing its transcriptional activation.

Selective proteolysis of ΔN variants may render cells more susceptible to cell death in response to stress or DNA damage, by shifting the equilibrium toward the TA proapoptotic proteins of the family. Interestingly, following DNA damage, ΔNTP73 and ΔNTP63 are rapidly degraded in an ITCH-independent manner, thus implying the existence of an E3, which specifically controls their degradation under these circumstances. While most of E3s targeting TP73 do not display any selectivity for the two major variants (TA and ΔN), the RING finger domain E3, PIR2/RNF144B provides specificity toward ΔNTP73 [Sayan et al., 2010]. The different affinity of PIR2/RNF144B for the oncogenic isoforms could be relevant to finely tuning the critical balance between the pro- and antiapoptotic isoforms in response to cellular stress. Additionally, PIR2/RNF144B is a direct transcriptional target of TATP73, suggesting the existence of a positive regulatory feedback loop. Stabilization of TATP73 following DNA damage would indeed sustain the apoptotic response by inducing PIR2/RNF144B expression that, in turn, favors ΔNTP73 degradation and activation of proapoptotic targets. Unlike TP73, both TATP63 and ΔNTP63 are targeted for ubiquitin-mediated degradation by PIR2/RNF144B.

Degradation of ΔNTP63 in response to genotoxic agents is promoted by nuclear factor κB (NF-κβ) in an IκB kinase-β (IKKβ)-dependent manner [Sen et al., 2010]. On the contrary, NF-κB has been shown to positively regulate TATP63 expression [Wu et al., 2010]. Similarly to the PIR2/Rnf144B/TP73 circuit, TATP63 positively regulates NF–κB transcription and protein stability by repressing its ubiquitin-mediated degradation [Sen et al., 2011]. The establishment of a regulatory feedback loop between TP63 and NF-κB ultimately leads to cell death by lowering and increasing ΔNTP63 and TATP63 cellular levels, respectively.

Following DNA damage and differentiation-inducing stimuli, ΔNTP63 is also degraded by FBW7, an E3 that interacts with TP53. A prerequisite for FBW7-triggered proteolysis of ΔNTP63 is binding to MDM2 in the nucleus, and its subsequent translocation to the cytoplasm [Galli et al., 2010]. Glycogen synthase kinase 3, which phosphorylates most of the known substrates of FBW7, seems to be required for ΔNTP63 degradation by FBW7. ΔNTP63 degradation in response to cellular stress is also partially dependent on RACK1. RACK1 is a WD-40 repeat-containing protein and a component of an SCF-like E3 complex [Li et al., 2009]. Mechanisms of selective destruction of ΔNTP63 might also be relevant during the elaboration and maturation of stem progenitor cells, which are exposed to extracellular signaling molecules, such as cytokine and growth factors that could trigger its destruction.

TP63 and TP73 degradation is further controlled by noncanonical ubiquitin-independent pathways. Genotoxic stress-mediated ΔNTP73 degradation can occur through the polyamine-induced antizyme pathway in a c-JUN-dependent manner [Dulloo et al., 2010]. Other proteasome-mediated pathways of TP63/TP73 degradation involve the NAD(P)H quinone oxidoreductase 1 [Asher et al., 2005], caspase [Ratovitski et al., 2001] and calpain cleavage [Munarriz et al., 2005], UFD2a [Hosoda et al., 2005], and cyclin G [Ohtsuka et al., 2003].

Transcriptional Activation Control

Modulation of TP63/TP73 steady-state protein levels by the ubiquitin/proteasome system represents one of the major regulatory mechanisms to modulate their transcriptional activation. Nevertheless, their function is further controlled by a number of protein interactors. These include enzymes that catalyze posttranslational modifications of TP63/TP73 and transcriptional coactivators or corepressors. Interestingly, many of these interactors can also act by influencing TP63/TP73 protein stability through distinct mechanisms.

A complex pattern of posttranslational modifications modulates TP63/TP73 biological activities in response to cellular stress. Following genotoxic stress, kinases are relevant regulators of the TP53 family. TP53 and its relatives share most of the kinases identified so far. The nonreceptor tyrosine kinase c-ABL, which is activated in response to DNA double-strand breaks generated by ionizing radiation or drugs, has been shown to control both TP63 and TP73 functions. Several reports highlight the importance of the TATP73/c-ABL functional interaction in enhancing the apoptotic response to DNA-damaging agents [Agami et al., 1999; Gong et al., 1999]. This is, at least in part, accomplished through protein stabilization of TP73 by c-ABL [Sanchez-Prieto et al., 2002; Tsai and Yuan, 2003]. c-ABL recruits TP73 through interaction of its SH3 domain with the PY motif of TP73, and catalyzes phosphorylation of TP73 at Tyr99 upon γ-irradiation [Agami et al., 1999]. Interestingly, modification of TP73 by c-ABL potentiates their interaction by stimulating the binding of tyrosine-phosphorylated TP73 to c-ABL SH2 domain [Tsai and Yuan, 2003].

In addition to exerting a direct control on TP73 activity, c-ABL orchestrates TP73 functions through phosphorylation-mediated activation of a number of other upstream regulators of TP73 or, alternatively, by enhancing their interaction with TP73. As an example, p38 MAP kinase (MAPK) is required for the activation of TP73 by c-ABL [Sanchez-Prieto et al., 2002]. p38 MAPK modifies threonine adjacent to proline residues on TP73. Inhibition of p38 MAPK blocks TP73 accumulation indicating that phosphorylation of TP73 by p38 MAPK controls TP73 protein levels [Bernassola et al., 2004].

In response to DNA damage stress, active c-ABL also phosphorylates the transcription coactivator YAP1 at Tyr357. Once tyrosine phosphorylated, YAP1 levels increase. Accumulated YAP1 influences TP73 function either by competing with ITCH for binding and thus allowing TP73 to escape proteasomal degradation [Levy et al., 2007], or by preferentially binding with TP73 to apoptotic gene promoters [Levy et al., 2008a]. YAP1 also favors the association of TP73 with the acetyltransferase p300 [Strano et al., 2005]. Modification of TP73 by p300 leads to protein stabilization [Bernassola et al., 2004] and confers promoter selectivity, with increased ability of TP73 to transactivate proapoptotic targets [Costanzo et al., 2002]. Following cellular stress, phosphorylation of TP73 by c-ABL augments its association with the prolyl isomerase PIN1 [Mantovani et al., 2004]. PIN1 binds phosphoproteins and modifies their conformation by isomerization of peptidyl–prolyl linkages. Modification of TP73 by PIN1 facilitates TP73 acetylation by p300 ultimately leading to protein stabilization and selection of promoter occupancy. All together, these regulatory mechanisms orchestrated by c-ABL contribute to TP73-elicited cell death in response to genotoxic stress.

Gonfloni et al. (2009) have reported that c-ABL regulates drug-induced TATP63 activation and cell death in oocytes. Oocytes represent the only known cellular system in which the intracellular steady-state concentration of TATP63α is kept high. Remarkably, in nonstressed oocytes, TATP63α is maintained in a closed inactive dimeric conformation, due to the presence of the TID. γ-irradiation-induced phosphorylation of TATP63 determines the formation of the transcriptionally active tetrameric conformation, which displays increased DNA-binding affinity [Deutsch et al., 2011]. c-ABL phosphorylates TATP63 on Tyr149, Tyr171, and Tyr289 residues. Interestingly, a phosphorylation-deficient mutant of TATP63 (Y149F) is unable to transactivate proapoptotic target genes, such as Puma and Noxa that mediate DNA damage-induced apoptosis. On the contrary c-ABL-catalyzed phosphorylation of ΔNTP63 has been reported to positively affect its protein stability [Yuan et al., 2010].

The ΔNTP63 is a target of the Homeodomain-interacting protein kinase 2 (HIPK2) in response to DNA-damaging insults [Lazzari et al., 2011]. HIPK2 has been reported to associate also with TATP73 and TATP63 isoforms [Kim et al., 2002], but the outcome of this interaction is still not fully understood. Phosphorylation of ΔNTP63 at Thr397 appears to be required for its degradation. Indeed, the nonphosphorylatable ΔNTP63α-T397A mutant is not susceptible to proteasomal proteolysis induced by drugs. Nevertheless, how HIPK2-mediated phosphorylation of TP63 influences its protein stability has not been clarified yet. HIPK2-induced phosphorylation of TP53 positively modulates its stability and sequence-specific DNA-binding activity through balancing its acetylation status [Puca et al., 2009]. Hence, one can speculate that HIPK2 may act similarly toward ΔNTP63. MacPartlin et al. (2005) reported that only TATP63 is able to associate with p300 in a DNA damage-dependent fashion, and that the acetylation activity of p300 is required for stimulating TATP63 transcriptional activity on the p21 promoter. Acetylation of ΔNTP63 by P300/CBP-associated factor (PCAF), that appears to increase with high cell confluency rather than in response to stress, promotes its nuclear export that is paralleled by p21 induction and cell cycle arrest upon cell contact [Chae et al., 2012]. Thus, even though ΔNTP63 modification by PCAF is not triggered by genotoxic stress, it seems to promote its inactivation through cytoplasmic sequestration and may likely favor its protein degradation.

Another kinase linking TP63 to protein acetylation by p300 is IKKβ. IKKβ-catalyzed phosphorylation of TATP63 inhibits its transcriptional activity by impairing the association with p300 [Liao et al., 2013].

A central regulator of all the TP53 family members is the promyelocytic leukemia protein (PML) [Salomoni et al., 2012] that recruits TP53, TP63, and TP73 to nuclear subdomains, named PML–nuclear bodies (NBs) [Guo et al., 2000; Bernassola et al., 2004; Bernassola et al., 2005]. Within these domains, PML facilitates the association of the TP53 proteins with common regulators and transcriptional coactivators, including HIPK2, YAP1, p300, and DAXX [Kim et al., 2003; Bernassola et al., 2004; Lapi et al., 2008]. Overall, the interactions taking place in the PML–NBs lead to protein modification of the TP53 proteins, which ultimately affect their stabilization and transactivation of proapoptotic targets. Thus, PML coordinates a number of crucial regulators of TP63/TP73 and establishes cross-talks among them to sustain their transcriptional function in response to stress. Under these conditions, PML is markedly upregulated, and the number and size of the PML–NBs increase. Interestingly, accumulation of TP73 in the NBs and its association with YAP1 is followed by transcriptional activation of the pml promoter by the TP73/YAP1 complex that, in turn, contributes to PML accumulation and NB formation [Lapi et al., 2008].

Conjugation of TP63/TP73 with small ubiquitin modifier-1 (SUMO-1) adds a further level of regulation of their transcriptional activities [Munarriz et al., 2004]. PIAS-1-mediated sumoylation of TP73 determines a reduction of its transcriptional activity on several target promoters, including p21, which in turn results in an increased number of cycling cells [Munarriz et al., 2004]. TP63 sumoylation is mediated by the SUMO-1-conjugating enzyme, Ubc9, and occurs at the end of the C-terminus. The transactivation abilities of TP63 proteins harboring mutations within the SUMO-1 motifs are enhanced compared with the wild-type counterpart, thus indicating that sumoylation represses TP63-dependent transcription [Huang et al., 2004]. According to other reports, SUMO-1 attachment to TP63 promotes its proteasomal degradation and sumoylation mutants display increased transcriptional potential as compared with wild-type TP63 [Ghioni et al., 2005; Straub et al., 2010]. All together, these findings demonstrate that sumoylation negatively affects TP63/TP73-driven transcription, and suggest that functional inhibition might be accomplished by decreasing their intracellular concentrations.

miRs Targeted by TP63

miRs represent a functional relevant class of TP63/TP73 target genes in mediating their biological activities including those related to tumorigenesis. miRs form one family of small noncoding regulatory RNAs. Several observations have implicated miRs in the initiation and progression of cancer [reviewed in Kasinski and Slack, 2011]. TP63 is extensively regulated by, and acts as a regulator of miRs in tumor cells. The important link between miRs targeted by TP63 and tumorigenesis is well-illustrated by Su et al. (2010) that report a role for TP63 as a master regulator of metastasis through the transcriptional regulation of Dicer and miR-130b. Rivetti di Val Cervo et al. (2012) found that ΔNTP63α inhibits senescence-specific miRs (miR-138, miR-181a, miR-181b, and miR-130b) by binding directly to TP63-responsive elements located in close proximity to their genomic loci. This pathway clearly contributes to the antisenescence function of TP63 and its deregulation could also be linked to cancer formation. The role of ΔNTP63 and miRs in tumor formation, instead, is well documented in HNSCC cells by Huang et al. (2013). They showed that HNSCC cells exposed to cisplatin displayed a dramatic ATM-dependent phosphorylation of ΔNTP63α that leads to the transcriptional regulation of downstream miRs, including miR-181a, miR-519a, and miR-374a (downregulated) and miR-630 (upregulated). In this study, the authors also report that ΔNTP63 binds to the Dicer promoter. Furthermore, TP63, through the transcriptional control of the miR-200 family and miR-205 also controls the epithelial-mesenchymal transition (EMT) [Gandellini et al., 2012]. The miR-200 family plays an important role in the regulation of EMT and tumor progression. This is achieved by directly repressing the transcription factors ZEB1 and ZEB2 [Gregory et al., 2008], which are central mediators of the EMT. While TP53 has been implicated in the regulation of miR-200c, ΔNTP63 specifically promotes miR-205 transcription in bladder tumors [Tran et al., 2013]. The TP63–miR-205 pathway has been also validated in prostate cancer cells, in which loss of TP63 and its transcriptional target miR-205 enhance metastasis in vivo [Tucci et al., 2012].

The Cross-Talk Between TP73 and miRs in Cancer

The first evidence of a possible cross-talk, although indirect, between TP73 and miRs has been shown by Sampath et al. (2009). They reported that treatment of chronic lymphocytic leukemia (CLL) with histone deacetylase inhibitors (HDACis) induces apoptosis in a TP73-dependent manner. Indeed, HDACis promotes the expression of miR-106b that, in turn, targets and inactivates ITCH. Decreased ITCH levels then lead to accumulation of the proapoptotic TATP73 isoform, and subsequent transcriptional activation of Puma and mitochondrial dysfunction. More interestingly, an inverse correlation between miR-106b and ITCH in CLL treated with the HDACi, LBH589 was reported. However, ITCH is also cleaved by caspases in CLL cells during apoptosis induced by various stimuli [Rossi et al., 2009]. This was further confirmed by Rivetti di Val Cervo et al. (2009), who reported the cleavage of ITCH due to the activity of caspase-3, -6, and -7 on residue Asp240. Moreover, caspase inhibitors were able to completely abrogate ITCH downregulation after HDACi treatment, suggesting that its downregulation is a consequence, rather than a cause, of apoptosis. Although, the involvement of TP73 in CLL was confirmed, the authors failed to validate ITCH as a target of miR-106b. This discrepancy could be explained by context-dependent action of miR-106b in CLL, as already described by several observations on other miRs [Fabbri et al., 2007].

Inactivation of TP53 in SCC is associated with both drug resistance and poorer prognosis; therefore, the identification of pathways mediating chemosensitivity could inspire novel therapeutic approaches. TP63 and TP73 are both overexpressed in SCC and they directly regulate the expression of miR-193a. In particular, ΔNTP63α, which is the predominant isoform expressed in the majority of human SSC [Thurfjell et al., 2005], acts as a transcriptional repressor of miR-193a, whereas TATP73 is a transcriptional activator of this miR. Interestingly, miR-193b can bind the 3′UTR of TP73, thereby inhibiting the expression of TATP73 [Ory et al., 2011]. This feedback, linking miR-193b to ΔNTP63 and TATP73, could explain, at least in part, the association between the expression of TP63 and chemosensitivity [Zangen et al., 2005]. Moreover, this pathway could also shed light on the limited outcome of chemotherapy in SCC. Indeed, cisplatin treatment leads to the degradation of ΔNTP63 and upregulation of TATP73, which, by inducing the expression of miR-193b, inhibits itself, thus negatively affecting the therapeutic intervention. These findings led the authors to speculate that miR-193b could be used as a potential biomarker in SCC for predicting the chemoresponse. Indeed, primary HNSCC display high levels of miR-193b. However, the small number of HNSCC cases analyzed so far limits the possibility of drawing definitive conclusions.

By using an integrative computational approach, Knouf et al. (2012) identified TP53 family members as regulators of miRs overexpression in ovarian carcinoma cells. Particularly, the expression of TP63 and TP73 significantly correlated with the expression of 17 miRs, whose promoter contains the binding sites for the TP53 family. Among those miRs, the expression of the miR-200 family shows positive correlation with that of TP73. Interestingly, there is no correlation between the TP53 mutational status and the expression of the miR-200 family, suggesting that TP53 is not the primary regulator of these miRs in ovarian cancer cells. Although these findings have been experimentally validated by luciferase and chromatin immunoprecipitation assays, indicating that TP73 directly binds to the miR-200 promoters, the biological significance of this interaction has not deeply investigated yet. Whether this correlation is associated with better prognosis or increased sensitivity to chemotherapy is still an open question. Moreover, while the axis TP63/miR in metastasis and tumor progression is well-documented [Su et al., 2010; Tucci et al., 2012], the role of TP73 in this biological context remains to be elucidated. However, a possible implication of TP73 in metastasis has been recently shown [Alla et al., 2012].

Transcriptional Targets Regulating Cellular Metabolism

The tumor-suppressor TP53 plays an important role in the regulation of cell metabolism. Indeed, TP53 regulates energy production promoting oxidative phosphorylation and dampening glycolysis. TP53 can also modulate autophagy. Moreover, oxidative stress can be either positively or negatively regulated by TP53 [Vousden and Ryan, 2009]. Several biological processes including cell proliferation, migration, senescence, and death can be mediated by reactive oxygen species (ROS).

Recently, the other family proteins have been also linked to cell metabolism control. A role for TATP73 in regulating aging through interference with mitochondrial metabolism has been elegantly showed by Rufini et al. (2012). TATP73 null mice display more pronounced aging as measured by several parameters including, decreased survival, weight loss, reduced body fat, corneal degeneration, and enhanced kyphosis. This phenotype is associated with increased oxidative damage and cellular senescence in the liver and kidney of the TATP73-null mice. Moreover, TATP73 depletion reduces cellular ATP levels, oxygen consumption, and mitochondrial complex IV activity, indicating a defect in mitochondrial metabolism [Rufini et al., 2012]. The expression of mitochondrial complex IV subunit cytochrome C oxidase subunit 4 (COX4i1), a protein essential for the assembly of fully functional mitochondrial complex IV, is decreased in TATP73-deficient mice, and restoration of its expression in TATP73 knockout cells is capable to rescue the reduced mitochondrial oxygen consumption. The increase in oxidative stress observed in the TATP73 null mice could drive the accumulation of oncogenic mutations that, in association with the previously described genomic instability [Tomasini et al., 2008], could account for augmented incidence of spontaneous tumors observed in these mice (Fig. 3; oxidative phosphorylation).

Metabolic reprogramming is a common feature of cancer cells. Indeed, rapidly proliferating cancer cells need to adapt their metabolism by increasing nutrient uptake and reorganizing metabolic fluxes to support biosynthesis [Cairns et al., 2011; Munoz-Pinedo et al., 2012]. An interesting and provocative aspect of TP73, as promoter of cell proliferation has recently emerged [Du et al., 2013]. The expression of glucose-6-phosphate dehydrogenase (G6PD), the rate-limiting enzyme of the pentose phosphate pathway (PPP) is under the direct control of TATP73. Interestingly, TP53 or TP63 are not able to bind the promoter and drive the expression of G6PD, suggesting that some cofactors that specifically interact with TATP73 are required for the regulation of this enzyme. The positive regulation of G6PD expression by TATP73 increases PPP flux and directs glucose to the production of NADPH and ribose, for the synthesis of macromolecules and detoxification of ROS. Overall, these findings suggest a possible oncogenic function for TATP73. However, how this new aspect of TATP73 can reconcile with its tumor-suppressor function is left hanging by the authors. One possible explanation is that TATP73 counteracts cellular senescence by activating an antisenescence metabolic response (Fig. 3; PPP pathway).

Cancer cells rely on specific metabolites to sustain proliferation and survival. In particular, serine pathway provides amino acids, glutathione (GSH), lipids, and nucleotides to the cell [Nomura et al., 2010; Cook et al., 2012]. Recently, TP53 has been associated to the capacity of cancer cells to face serine starvation. Indeed, cells lacking TP53 failed to respond to serine starvation, due to oxidative stress condition, which leads to reduce viability and severely impaired proliferation [Maddocks et al., 2013].

Interestingly, Amelio et al. (2013) found that there is a correlation between serine biosynthetic pathway and TP73 expression in human lung adenocarcinomas, the predominant tumor developed by the TATP73 knockout mice. Moreover, metabolic profiling of human cancer cell lines revealed that TATP73 activates serine biosynthesis, resulting in increased intracellular levels of serine and glycine, associated to accumulation of glutamate, TCA anaplerotic intermediates, and GSH. Mechanistically, TATP73 directly regulates the expression of glutaminase-2, GLS2, which promotes the conversion of glutamine in glutamate, and, in turn, drives the serine biosynthetic pathway. More importantly, inhibition of TATP73 expression completely abrogates cancer cell proliferation upon serine/glycine deprivation, suggesting a role for TP73 in supporting cancer cells under metabolic stress. Therefore, TP73 and TP53 together appear to help cancer cells to face the oxidative stress associated with serine starvation. Pharmacological inhibition of the enzymes involved in serine biosynthesis may represent a novel therapeutic approach for cancers harboring TP53 and/or TP73 loss of function (Fig. 3; glutaminolysis). Cell fate is regulated by distinct processes [Melino et al., 2005; Denton et al., 2012], including caspase-mediated apoptosis [Inoue et al., 2009], necroptosis [Kroemer et al., 2009; Vanlangenakker et al., 2012], and autophagy [Denton et al., 2012].

Autophagy is a catabolic process that plays a pivotal role in maintaining cellular homeostasis by monitoring the integrity of long-lived proteins and organelles [Levine and Klionsky, 2004]. Both tumor suppression and oncogenic functions have been ascribed to autophagy [Kondo et al., 2005; Maiuri et al., 2009]. Indeed, at the early stage of tumor development, autophagy acts as a tumor-suppressor mechanism, whereas at more advanced stages, it promotes tumor progression [White, 2012]. All the members of the TP53 family transcriptionally activate several autophagy-related genes, including Atg2b Atg4a, Atg4c, Atg5, Atg7, Atg10, Tmem49, Ulk1, Ulk2, and Uvrag [Kenzelmann Broz et al., 2013]. The upregulation of these genes results in the induction of autophagy in various contexts, including in response to DNA damage, and after genetic activation. This was shown in primary mouse embryonic fibroblasts and lung cancer cell lines. Overall, in this context, the activation of TP53 family does not promote survival but is required to sustain TP53-dependent apoptosis and suppression of cell transformation. This finding reinforces the role of TP73 in autophagy as previously demonstrated. Indeed, TP73 is able to regulate the expression of damage-regulated autophagy modulator, DRAM1 [Crighton et al., 2007]. In addition, TP73 levels are increased after treatment of cells with rapamycin, an inhibitor of mTOR signaling, suggesting a cross-talk between mTOR and TP73 in autophagy and metabolic pathways [Rosenbluth et al., 2008]. The involvement of TP73 in the autophagic process has been recently demonstrated in vivo as well [He et al., 2013]. When TP73-deficient mice are subjected to nutrient deprivation conditions, there is an extensive accumulation of lipid droplets in the liver, paralleled by reduced levels of autophagy. Therefore, it is reasonable to speculate that triglyceride hydrolysis conversion into fatty acids is blocked resulting in deficient autophagy. According to this in vivo phenotype, the expression of Atg5, a gene essential for autophagosome formation, is reduced in several organs including liver, colon, and heart of TP73 deficient mice. Mechanistically, the authors proved that Atg5 is a direct transcriptional target of TP73. In particular, TATP73α and TATP73β bind and transactivate the Atg5 promoter. The TP73-ATG5 axis further confirms the role of TP73 in the regulation of autophagy and moreover, contributes to the explanation of why the TP73-deficient mice show the complex phenotype described above. In fact, a role of autophagy during neuronal development and immune response has been described [Mizushima and Levine, 2010].

Similarly, TP63 controls some aspects of cellular metabolism, which could contribute to its tumor-suppression function. In particular, TATP63 plays a critical role in regulating energy metabolism. Indeed, TATP63 selective knockout mice develop obesity, glucose intolerance, and insulin resistance [Su et al., 2012]. Moreover, these mice show defects in fatty acid oxidation and display mitochondrial dysfunction. This phenotype is mechanistically explained by deregulated expression of metabolic enzymes, including fatty acid synthase and carnitine palmitoyltransferase (CPT-1), in the liver of the knockout mice, that results in increased fatty acid synthesis and reduced fatty acid consumption. A deeper analysis showed that TATP63 transcriptionally regulates the expression of SIRT1, AMPKα2, and LKB1, which are well-known regulators of the enzymes involved in lipid metabolism. More importantly, the reexpression of these target genes is sufficient to rescue some of the metabolic defects observed in the TATP63 knockout mice, indicating that TATP63 regulates energy metabolism through SIRT1, AMPKα2, and LKB1.

As mentioned above, GLS2 is the enzyme responsible for the conversion of glutamine into glutamate, which, in turn, can be converted into α-ketoglutarate, thus stimulating the production of ATP via the TCA cycle. In addition, glutamate can also be used within the cell for the production of GSH. Enhanced mitochondrial glutamine catabolism is one of the major metabolic alterations found in cancer cells [Cairns et al., 2011]. In a cohort of human primary colon cancers, the expression of GLS2 was found significantly upregulated and was accompanied by a concomitant upregulation of TP63 levels, suggesting a possible role of the TP63/GLS2 axis in the maintenance of glutamine metabolism in colon cancer cells. The connection between TP63 and GLS2 in tumors is also reinforced by the fact that GLS2 can be directly [Giacobbe et al., 2013], or indirectly via SIRT1 [Su et al., 2012], regulated at transcriptional level by TATP63.