P63 and P73 Activation in Cancers with p53 Mutation

The members of the p53 family comprise p53, p63, and p73, and full-length isoforms of the p53 family have a tumor suppressor function. However, p53, but not p63 or p73, has a high mutation rate in cancers causing it to lose its tumor suppressor function. The top and second-most prevalent p53 mutations are missense and nonsense mutations, respectively. In this review, we discuss possible drug therapies for nonsense mutation and a missense mutation in p53. p63 and p73 activators may be able to replace mutant p53 and act as anti-cancer drugs. Herein, these p63 and p73 activators are summarized and how to improve these activator responses, particularly focusing on p53 gain-of-function mutants, is discussed.


Introduction to the p53 Family
The p53 family has three members, p53, p63, and p73 [1][2][3]. TA (transactivation) isoforms of p53 family members are tumor suppressor genes [4,5]. p53 has a high frequency of mutation in cancers causing loss of its tumor suppression function [6,7]; however, p63 and p73 are rarely mutated in cancers [8][9][10]. In this review, we briefly introduce each of the members with an emphasis on the most common mutations of p53 making it nonfunctional. Further, we summarize p63 and p73 activators that can replace them to obtain a similar tumor suppressor function in the p53 family. Some p53 mutants can obtain oncogenic function as a gain of function similar to an oncogene [11,12]. We also discuss how to improve p63 or p73 activator drug response in p53 gain-of-function mutation cancer cells.

p53
p53 was the second tumor suppressor gene identified, although p53 was actually discovered in 1979, before the first tumor suppressor gene Rb, which was cloned in 1986 [13]. (A) p53 has, respectively, a 10.72% and 62.74% nonsense mutation rate and missense mutation rate in all cancer mutation samples from the COSMIC database (https://cancer.sanger.ac.uk/cosmic ; accessed date (29 April 2022) [18]. (B) p63 has, respectively, 2.97% and 32.34% nonsense mutation rate and missense mutation rate in all cancer mutation samples. (C) p73 has, respectively, 2.28% and 27.33%, nonsense mutation rate and missense mutation rate in all cancer mutation samples.

Nonsense Mutation
p53 nonsense mutations comprise ~10% of all p53 mutants ( Figure 1A); the actual p53 nonsense mutation rate is higher than the average ~5% [38]. There are three pre-stop DNA codons, TAA, TAG, and TGA. Nonsense mutation leads to the generation of premature termination codons (PTC), which leads to nonsense-mediated mRNA decay (NMD), resulting in the inability to express full-length proteins and extremely low expression levels Figure 1. p53 has a high nonsense and missense mutation rate compared to p63 and p73 in cancer. (A) p53 has, respectively, a 10.72% and 62.74% nonsense mutation rate and missense mutation rate in all cancer mutation samples from the COSMIC database (https://cancer.sanger.ac.uk/cosmic; accessed date (29 April 2022) [18]. (B) p63 has, respectively, 2.97% and 32.34% nonsense mutation rate and missense mutation rate in all cancer mutation samples. (C) p73 has, respectively, 2.28% and 27.33%, nonsense mutation rate and missense mutation rate in all cancer mutation samples.

p63
p63 was cloned in 1998 [27]. The major C-terminal p63 isoforms are p63α, while dominant-negative ∆Np63 was the predominant N-terminal isoform in most tissues from the p73-High/p63-High group [28]. p63 is a rare rarely mutated in cancers [8]. According to the COSMIC database [18], there are 55,869 unique clinical samples with 2160 unique samples having p63 mutations ( Figure 1B). The nonsense mutation rate of p63 is only 2.97%, and the missense mutation rate of p63 is 32.34%. Although p63 is rarely mutated in cancers, distinct p63 germline mutation can cause several different types of abnormal development issues. Ectrodactyly, ectodermal dysplasia, and cleft lip/palate (EEC) syndrome are mainly characterized by severe ectrodactyly and limb defects with a p63 missense mutation in the middle of the DNA binding domain [29]. The distinguishing features of ankyloblepharon-ectodermal defects-cleft lip/palate (AEC) syndrome are ankyloblepharon, congenital erythroderma, skin fragility, atrophy, palmoplantar hyperkeratosis, and extensive skin erosions with p63 missense mutation in the C-terminal the sterile-α-motif (SAM) domain and transcriptional inhibitory (TI) domains [30,31]. Different p63 mutations can also cause isolated split hand/foot malformation (SHFM) nonsyndromic diseases with a missense mutation of K193E and K194E and nonsense mutants of Q634X and E639X on TAp63α [32]. p63 knockout mice have been found to fail to form the stratified epidermis, limbs, teeth, mammary glands, and postnatal lethality due to dehydration [33,34].

p73
p73 was cloned in 1997 [35]. Compared to p63, p73 also commonly expresses α isoforms [28] and the expression of TAp73 is higher than ∆Np73 [36]. p73 is rarely mutated in cancers [9,10]. According to the COSMIC database [18], there are 42,580 unique clinical samples with 879 unique samples with p73 mutations ( Figure 1C). The nonsense mutation rate of p73 is only 2.28%, and the missense mutation rate of p73 is 27.33%. There are no reports about p73 germline mutation in relation to any type of genetic disorder or syndrome. But hydrocephalus, hippocampal dysgenesis, and pheromonal defects can be found in p73 knockout mice [37].

Nonsense Mutation
p53 nonsense mutations comprise~10% of all p53 mutants ( Figure 1A); the actual p53 nonsense mutation rate is higher than the average~5% [38]. There are three pre-stop DNA codons, TAA, TAG, and TGA. Nonsense mutation leads to the generation of premature termination codons (PTC), which leads to nonsense-mediated mRNA decay (NMD), resulting in the inability to express full-length proteins and extremely low expression levels of truncated proteins [39]. Two mechanisms are known to regenerate full-length proteins, one is to inhibit NMD, and the other is for PTC readthrough. It is known that aminoglycoside drugs, such as G418 and gentamicin, can inhibit NMD and promote p53 PTC readthrough [40,41]. But these drugs are highly toxic and cannot be used in clinical practice. 2,6-Diaminopurine (DAP), can inhibit the activity of putative ribosomal RNA methyltransferase 1 (FTSJ1) to increase the capacity of tRNATrp to recognize the UGA stop codon to promote p53 PTC readthrough, but this drug does not have the ability to inhibit NMD [42]. Furthermore, DAP is only effective for nonsense mutations of TGA but not TAA or TAG [42], and this greatly reduces the available targets. In addition, some phthalimide derivatives and antimalarial drug quinines can promote the p53 PTC read-through ability of G418 to increase the proportion of full-length p53 and to reduce the expression of the truncated protein, but these drugs alone have no effect on PTC read-through [43,44]. Nonaminoglycoside drugs, such as Ataluren (PTC124), also increase the read-through ability of PTC without the ability to inhibit NMD, and PTC124 has been used in clinical phase II or III trials to treat genetic diseases with specific nonsense mutations [45,46]. PTC124 can also promote p53 PTC readthrough [47]. A recent study has shown that CC-885 and CC-90009 can inhibit NMD; of note, the effective concentration of CC-885 for treatment of p53 nonsense mutation with TAA is only one-tenth of that of CC-90009 [48].

Loss-of-Function Mutants
p53 missense mutations contain both loss of function and gain of function. p53 mutation is mainly located at the N-terminal transactivation domain or middle DNA binding domain [2]. In addition, several point mutants still have normal DNA binding function [49][50][51][52]; most p53 mutations within the N-terminal transactivation domain or the DNA binding domain lose their transactivation function or DNA binding function causing loss of their tumor suppressor functions such as cell cycle checkpoint controls and apoptosis [53][54][55]. These p53 mutations can associate with p63 and p73, whereas wild-type p53 cannot [56,57]. Therefore, loss-of-function p53 mutants act in the same way as the ∆N isofroms of the p53 family having a dominant-negative effect to repress the functions of normal TA isoforms of p53 family members [57,58].

Gain-of-Function Mutations
Some p53 mutants can obtain some oncogenic functions such as cell migration, invasion, and metastasis to enhance tumorigenesis [59], and these p53 mutants are called gain-of-function mutants. The acquisition of p53 gain-of-function is via three mechanisms [60,61]. First, mutant p53 can directly bind to the novel binding site with a p53 non-canonical sequence to activate several oncogenic genes [51]. Second, mutant p53 can act as a co-activator to bind to other transcription factors to activate some oncogenic genes [62]. Third, mutant p53 can bind to other tumor suppressive-type transcription factors to cause loss of transcription ability [63]. Some p53 mutants can become aggregated in several types of cancer, such as breast, lung ovary, colorectal, and head and neck cancers [64][65][66][67]. It is known that these aggregations of mutant p53 can sequester other tumor suppressor genes as a third mechanism to cause p53 gain of function [68][69][70].

p63 Activation Drugs
Because TAp63 isoforms have limited expression in organs [28], only a few reports have looked at TAp63 activation for anti-cancer purposes; they are summarized in Table  1. Bliotoxin was able to upregulate the levels of DAPK1 to induce TAp63 but not p53 or TAp73 expression to induce apoptosis in paclitaxel pretreated paclitaxel-resistant CaOV-3 and SK-OV-3 ovarian cancer cells [96]. Lovastatin was able to induce TP63 transactivation through phosphorylation of the AMPK-p38MAPK-TAp63 cascade to cause hypopha-

p63 Activation Drugs
Because TAp63 isoforms have limited expression in organs [28], only a few reports have looked at TAp63 activation for anti-cancer purposes; they are summarized in Table 1. Bliotoxin was able to upregulate the levels of DAPK1 to induce TAp63 but not p53 or TAp73 expression to induce apoptosis in paclitaxel pretreated paclitaxel-resistant CaOV-3 and SK-OV-3 ovarian cancer cells [96]. Lovastatin was able to induce TP63 transactivation through phosphorylation of the AMPK-p38MAPK-TAp63 cascade to cause hypopharyngeal carcinoma FaDu cell death [97]. TAp63 can active PUMA (p53 upregulated modulator of apoptosis) [98,99], and interferon-α can induce TP63 and PUMA expression in hepatocyte derived cellular carcinoma cell line HuH7 cells [100]. miR-130b mimics can activate TAp63 and repress ∆Np63 to decrease cell viability in ovarian cancer Ovcar1-8 cells [101]. miR-124 mimics can activate TAp63 and repress ∆Np63 to inhibit cell growth in LoVo and SW480 colorectal cancer cells [102]. miR-140 can directly target p63 3 -UTR sequences to repress p63, and the miR-140 inhibitor can activate TAp63 in HGC-27 and BGC-823 gastric cancer cells to induce cell apoptosis [103].

Influence of Interactions between p53 Family Members on p63 or p73 Activators and the Importance of Combination Treatment Strategy
Several factors can influence p63 or p73 activation in p53 mutant cells. First, a high level of dominant-negative ∆Np63 or ∆Np73 expression in cells can block the TAp63 or TAp73 activity [57,58]. Second, gain-of-function mutant p53 can become aggregated and co-aggregate with TAp63 or TAp73 [68,133]. miR-130b and miR-124 mimics not only induce TA isoform p63 in p53 mutant cells, but also repress the relative proportion of ∆N isoform p63 to resolve the first issue [101,102]. Prodigiosin, metformin, diallyl disulfide, and MEK1 inhibitors not only induce TA isoform p73 in p53 mutant cells, but also repress the relative proportion of ∆N isoform p63 or p73 [107,110,115,124,134]. Recently we found that p73 activators RETEA and NSC59984 have a poor response in aggregative p53 mutant HNSCC cells compared to non-aggregative p53 mutant cells. Furthermore, using an NAMPT inhibitor to block p53 aggregation can enhance the anti-cancer effect of p73 activator RETEA and NSC59984 in p53 gain-of-function mutant HNSCC cells [135]. This combination treatment strategy comprised of a p73 activator and a p53 aggregation inhibitor was able to resolve the second issue to active p73 in p53 gain-of-function mutants. Because there are fewer p63 activators than p73 activators (Tables 1 and 2), whether this combination treatment strategy with a p63 activator and p53 aggregation inhibitor is workable or not in p53 gain-of-function mutants still needs further investigation. However, it is likely that co-treatment with a p53 aggregation inhibitor can improve p63 and p73 drug response in p53 gain-of-function mutants.

Viral Proteins and the p53 Family
Some cells contain wild-type p53, but several of these cell lines may be infected by certain viruses to express viral protein(s) that interact with p53 family members to cause their loss of function. Hela cells were infected by oncogenic HPV type 16 (HPV16) with E6 oncoproteins to induce ubiquitin-dependent proteolytic degradation of wild-type p53 [136,137]. HPV18 E6 can interact with TAp63β but not the other p63 isoforms to induce degradation of the TAp63β [138]. HPV18 E6 can also interact with TAp73α or TAp73β to reduce its transcriptional activity but has no influence on p73 stability. Diallyl disulfide can still induce p73 to mediate the cell apoptotic program in Hela cells [115]. Other oncogenic viral proteins like the LANA protein of the Kaposi sarcoma virus, the BZLF1 protein of the Epstein-Barr virus, and the papain-like protease of the nonstructural protein 3 of SARS-CoV also can cause p53 degradation [139][140][141]. In addition, the HbX protein of the Hepatitis virus B and the NS2 proteins of hepatitis virus C can interact with p53 to reduce its transcriptional activity [142,143]. But whether all these oncogenic viral proteins can influence p63 and p73 or not still remains to be addressed. This means that p63 and p73 activators may also not be such powerful anti-cancer agents in p53 mutant cells infected by certain viruses. Interestingly, HPV18 E7 can inhibit the interaction of p53 and MDM2 to stabilize p53 to increase the transcriptional activation function of p53 [144]. The NS5 proteins of the Zika virus can interact with p53 to prolong the half-life of p53 to cause cell death [145]. Therefore, some viral protein(s) can also promote wild-type p53 function. Whether p63 and p73 can be activated by these viral protein(s) or not is also unknown.

Discussion and Concluding Remarks
There are also several p53 activators that can directly bind to mutant p53 to convert it into a wild-type like structure [154][155][156]. These drugs directly reactivate and change p53 activity. Such drugs were not emphasized in this review. For example, PRIMA-1 and its methylated form PRIMA-1Met (APR-246) can directly bind to mutant p53 to refold it as a wild-type p53-like conformation [157][158][159], because mutant p53 can associate with p63 and p73, whereas wild-type p53 cannot [56,57]. Therefore, PRIMA-1Met has also been reported to activate p63 and p73 through relieving them from mutant p53 that had been converted into wild-type like structure [160][161][162].
In conclusion, this review focused on summarizing the drugs that can activate p63 or p73 in p53 mutant cancer cells. Besides synthetic chemicals and purified natural products, synthesized small biomolecules, like microRNA mimics or inhibitors have also been used to activate p63 and p73 in p53 mutant cancer cells (Tables 1 and 2). It has also been pointed out in this review that viral proteins and p53 isoforms may also influence the drug response of p63 and p73 activators in p53 mutant cancers.