The Sabapathy lab is focused (i) on understanding the mechanistic basis of cellular transformation and resistance to cancer therapy, by investigating the key p53/p73 tumor-suppressor pathway, so as to generate novel therapeutic strategies; and (ii) on developing mouse models to best reflect various cancer types to study the carcinogenic process.
Cancer genome sequencing efforts have led to the identification of major alterations that occur in a wide variety of genes. Moreover, expression analyses have also revealed differences in many key genes that have been suggested to regulate the tumorigeneic process and response to therapy. Of these, some – especially those that code for enzymes - have been deemed “druggable” and hence, intensely pursued, albeit being of relevance to a small population. On the other hand, many genes have not received significant attention from the drug discovery perspective, simply due to the perception that they are “undruggable”. One such gene family is the p53 family of tumor suppressors, which are transcription factors. While p53 is the most mutated gene in cancers, its homologue p73 is often overexpressed in cancers. Large amount of data exist to demonstrate the role of mutant p53 and overexpressed p73 in tumor promotion and resistance to therapy. While these molecules could be potential therapeutic targets of enormous benefit to patients, they have not been subjected to major drug discovery efforts due to the difficulties in targeting transcription factors.
Mutant p53 – the real target for cancer therapy?
The recent massive whole-genome sequencing efforts have confirmed that p53 is the most mutated gene across all cancer types. Most of the mutations occur in the DNA-binding domain (DBD) of p53, thereby abolishing its function as a transcription factor, and thus, processes such as apoptosis and cell-cycle arrest. When mutant p53 co-exists with the wild-type (WT) allele in the early stages of cancer, mutant p53 exerts a dominant-negative (DN) effect over the WT allele, thereby attenuating the cell-death response to DNA-damaging agents (Fig.1a). However, in later stages of cancer development, when the WT allele is lost due to loss-of-heterozygosity, the remaining mutant p53 exerts gain-of novel functions that drive cellular invasion (Fig.1b), and thus, metastasis. In this respect, we have shown that cancer cells are addicted to mutant p53 for survival, and abolishing mutant p53 expression can lead to sensitization to death and thus growth retardation (Fig.1c). Similarly, the DN effect of p53 can be overcome by reducing the mutant p53 expression, as we had shown using a hypomorphic strain of mutant p53 knock-in mice (Fig.1a). These data therefore demonstrate that if mutant p53 levels can be regulated, it will to be an ideal target for therapeutic intervention. Over the years, we have performed systematic analyses to understand the functions of mutant p53, and have demonstrated that the even similar p53 mutants may have both common and differing properties, indicating that any targeting effort should be aimed at generating “specific-mutant-p53”-targteing molecules (rather than pan-mutant p53). In this aspect, we are embarking on a program to identify novel molecules that would target mutant p53 expression without having an effect on the WT allele, using a multitude of novel screening strategies.
Besides the DBD-mutations, how mutations in the other domains of p53 affect response to therapy and cancer development has not been understood fully. In this regard, we have identified that amino terminal-p53 (ATp53) mutations often lead to loss of full-length p53 expression, but concurrently lead to the expression of a truncated p53 form referred to as p47 (Fig.2a). Interestingly, p47 appears to be also capable of inducing apoptosis and inhibiting growth, even though it lacks the transactivation domain (Fig.2b). Nonetheless, ATp53 mutations are found both in sporadic and familial human cancers, indicating that apoptosis proficiency is insufficient for p53 to inhibit tumor formation. The mechanistic basis of p47’s functions in tumorigenesis and its ability to selectively induce apoptosis are being investigated.
Figure Legend: Figure 1. Reducing mutant p53 expression promotes cell death. (A). Thymocytes from mice carrying various p53 alleles were subjected to g-irradiation and cell death was determined (lower panel). Top panel shoes expression levels of p53, and the hypomorphic p53 mutant (Mutant-hypo) shows reduced expression compared to the mutant (upper panel). (B). Expression of mutant p53 (i.e. R175H) in p53 null cells lead to enhanced cellular migration in scratch assays after 20h. (C). Silencing expressing of endogenous mutant p53 leads to reduced cellular survival. Colonies after 2 weeks of silencing are shown.
Figure Legend: Figure 2. Amino-terminal p53 mutations. (A). Human cancer-derived amino-terminal p53 mutations do not express p53, but express a shorter p47 form (arrowheads). (B). Expression of p47 leads to inhibition of cellular growth as effectively as p53.
p73 – adding to the complexity
p73 is a homologue of p53 and belongs to the same family of tumor suppressors. It exists as 2 major forms: the full-length TAp73 from (similar to p53), and the amino-terminal truncated DNp73 (similar to p47) (Fig.3a). However, unlike p53, it is hardly mutated, but both forms are often variably over-expressed in many cancers, suggesting that p73 may have other differing roles in regulating tumorigenesis. While absence of TAp73 promotes tumor formation in mice – albeit weakly, highlighting a role in tumor suppression, its role in supporting tumor growth had been controversial. Our work over the years has focused on understanding this property, and we had earlier shown that TAp73 is capable of driving cellular proliferation in specific contexts, through the regulation of AP-1 target genes. Recently, we have demonstrated that TAp73 is capable of inducing the expression of angiogeneic genes, especially in hypoxic conditions that are prevalent in tumors. Thus, TAp73 appears to be utilizing multiple mechanisms to promote cancer cell growth, implying that these tumor-promoting pathways may be targetable for improving therapeutic response. We have therefore embarked on identifying key molecular determinants of the TAp73 pathway through high-throughput whole-genome siRNA screens and proteomics approaches, with the eventual goal of inhibiting them to reduce angiogenesis and proliferation of cancer cells. In addition, given the diametrically opposite roles of TAp73 in both promoting and suppressing tumor formation (Fig. 3b), we are now poised to address the question on how a transcription factor like TAp73 is able to regulate these seemingly opposite cell fate outcomes. We have therefore just embarked on answering this question of cell fate, using novel screening technology and next-generation animal models to study the spatial and temporal role of TAp73 both within the tumor exclusively, as well as in the stromal compartments.
In contrast to TAp73, the role of DNp73 in tumor promotion is somewhat expected due to its known ability to bind to both p53 and TAp73 and inhibit their functions. Incidentally, DNp73 overexpression in cancers also correlates with resistance to therapy and poor prognosis, suggesting that DNp73 could be targeted to improve therapy. While one can envisage that both the TAp73 and DNp73 forms may be similarly regulated, our results have shown the exposure to DNA-damaging agents lead to upregulation of TAp73, but to the down-regulation of DNp73, highlighting the complex nature of their regulation. We have thus worked out the mechanisms of DNp73 degradation, and found that it is NOT degraded by the ubiquitin pathway upon DNA-damage, but is degraded through the antizyme (Az) pathway, which also targets other oncogenes such as cyclinD1 and Aurora A (Fig.4). We have thus embarked on a program to identify molecules that target this pathway with a view of targeting the DNp73-overexpressing cancers.
Figure Legend: Figure 3. p73. (A). Schematic comparing p53 with TAp73 and DNp73. TA: transactivation; DBD: DND-binding domain; OD: oligomerization domain; SAM: sterile alpha motif domain. (B). Expression of TAp73 leads to growth inhibition (left), but it is overexpressed in many human cancers (right), highlighting a conundrum.
Figure Legend: Figure 4. Schematic shows the Antizyme (AZ) pathway, which is induced by frameshifting upon polyamine addition, and leads to target degradation (such as DNp73). AZI is the Az inhibitor.
Mouse models for hepatocellular carcinoma, liver fibrosis and liposarcoma
Another major effort in the laboratory is to develop mouse models that would recapitulate the human cancer conditions as best as possible, using state-of-the-art genetic engineering technology. This will enable the identification of novel biomarkers for early detection, as well as potential molecular targets for timely-intervention. In this regard, we have been working on modeling hepatocellular carcinoma (HCC), and liver fibrosis, and have generated mouse models that recapitulate human HCC, both molecularly and histologically (Fig.5a). Moreover, we have established the liver fibrosis model in mice, using carbon-tetrachloride, where the fibrotic symptoms could regress upon withdrawal of treatment (Fig.5b). We are now interrogating the critical aspects of the fibrotic process, through the analysis of the functions of several transcription factors that are major regulators of liver development and pathology, through their deletion in multiple cells types of the liver (Fig.5c).
Other efforts are also ongoing to establish mouse models for liposarcoma, tumors that arise from fat cells (adipocytes) in soft tissues. Though surgery is the main mode of treatment, the understanding of this disease is limited due it being not a common cancer, and hence, treatment modalities have been restricted. While the process of adipogenesis has been well studied, knowledge of the transformation of an adipocyte to liposarcomas is limiting due to the lack of effective model systems to study the development and progression of this disease. Mdm2, the negative regulator of the tumor suppressor p53, is often amplified in all types of sarcomas. Thus, we are generating mouse models in which selective genetic changes are introduced in the germ-line conditionally to follow the transformation of the adipocytes, which will provide a paradigm for studying the biology of liposarcomas, and could open up new opportunities for treatment.
Figure Legend: Figure 5. Mouse models for HCC and fibrosis. (A). Hepatisis (HbSAg) transgenic mice or wild-type (WT) mice were injected with aflatoxin-B1 or oil (control) and followed for 15 months to model the development of HCCs. Representative pictures of livers are shown at the various time points of harvest. (B). WT mice were injected with carbon-tetrachloride to induce fibrosis for 8 weeks, followed by recovery for 6 weeks. Pictures show Sirius staining and its quantification to determine extent of fibrosis. (C). Model to study liver diseases in various cell types of the liver, as indicated, using the respective cre-transgenic mice to delete the genes being studied. qHSC: quiescent hepatic stellate cells; aHSC: activated HSC.
Prospective students to apply through SINGA
and NUS (http://medicine.nus.edu.sg/postgrad/index.shtml)
PhD program scholarships.
Interested A*STAR scholars and post-doctoral fellows to contact the PI directly.