Prof. Varda Rotter, Research Activities

Varda Rotter
The Norman and Helen Asher Chair of Cancer Research
Department of Molecular Cell Biology
Weizmann Institute of Science
Rehovot, Israel
Prof. Varda Rotter
Tel: Lab: + 972-8-934-4501
Office: + 972-8-934-4071
Home: + 972-8-934-6498
Fax: + 972-8-946-5265, 972-8-934-4125

  • p53 in Normal and Tumor Cells

    It is well accepted that the p53 tumor suppressor gene plays a pivotal role in protecting cells from cancer development. Inactivation of p53, the "guardian" of the human genome, causes genetic instability leading to the accumulation of genetic alterations, which in time induce malignant transformation of cells. Several lines of research showed that mutant p53 protein, which frequently accumulates in human tumors (Rotter 1983), facilitate tumor initiation and progression by a “gain of function” mechanism. Research in our laboratory is focused on revealing the mechanism by which wild type p53 acts in normal cells and the way in which mutant p53 contributes to cancer development. Our working hypothesis is that mutant p53 activates specific target genes that turn on specific gene networks, which eventually lead cells to cancer (Sigal and Rotter 2000).

    Oncogenic funtion of mutant p53 proteins

    In our experiments we found that expression of mutant p53 acquires cells with resistance to chemotherapy (Li et al., 1989). Cells expression of mutant p53 with an intact N-terminal transcription domain are more resistant to drug induced apoptosis (Matas et al., 2001). In an effort to identify the mutant specific target genes we have adopted a genome-wide analysis approach. We used cDNA micro-array technology and compared the gene expression pattern of p53-null cells with their well-defined p53 mutant expressing counterpart cell lines that we have generated. This approach yielded lists of potential mutant p53 target genes. Interestingly, such targets are the EGR1 (Weisz et al., 2004), MST1 (Zalcenstein et al., 2006) and others. In our study we showed that mutant p53 indeed interacts with their promoter regions and that knocking-down these specific targets significantly reduced the resistance of cells to chemotherapeutic drugs. In independent studies, we observed that mutant p53 seem to affect TPA-induced ATF3 cell death (Buganim et al. 2006) and was found to modify the activity of TGFβ induced cell growth arrest (Kalo et al., 2007). We also found that mutant p53 interferes with AIF induced apoptosis (Stambolsky et al., 2006) and NFkb induced networks (Weisz et al. 2007a). In all, our results suggest that mutant p53 actively modifies central cell growth control networks (Weisz et al., 2007b).

    As it is well accepted that malignant transformation is a step-wise process, it challenges us to discover which of these events involve the p53 protein. Deciphering the molecular regulatory events that drive malignant transformation represents a major challenge for systems biology. We have consequently established several in vitro transformation models in which normal cells are transformed into cancer cells by well-controlled genetic alterations. In our experiments, we have immortalized various human primary cells of lung and prostate origin, and engineered several defined cancer-associated genetic alterations into them. These included inactivation of p53 tumor suppressors by several methods, over-expression of mutant p53, over-expression of the ras oncogene and various combinations of these genetic modifications. As a result, we have obtained transformed cells that are capable of developing tumors in mice. This suggests that the in vitro developed system represents an authentic model of cancer development. To evaluate the gene networks that are associated with the malignant defined steps, we have used a genome-wide approach, thereby permiting the identification of gene clusters that are associated with the individual steps of malignant transformation that we defined. By haven taken this approach we identified specific gene signatures that are associated with the specific genetic alterations along the process of cancer development. (Milyavsky et al., 2003).

  • We have focused in particular, on four important clusters that significantly represent this process. While establishing our in vitro system, we noticed that although cells immortalized by hTERT grow well in vitro, at a certain time point, varying between the various cells types studied, cells seem to undergo a certain cell crisis that is then followed by the emergence of cells with a faster cell growth rate. We refer to this transition point as a junction where “slow” cells are turned into “fast” growing ones. Analysis of that checkpoint revealed a loss of the p16 tumor suppressor gene. This concurs with other reports where it was shown that the p16 is undergoing methylation and thus is transcriptionally silenced. Gene cluster analysis indicated that at that time point there is a specific loss of gene clusters, consisting of typical cell differentiation (cluster 1) and a concomitant gain of a gene cluster that contains genes associated with cell proliferation (cluster 2). We confirmed that at that turnpoint the immortalized WI-38 human fibroblasts lost their ability to undergo cell differentiation following a TGFβ trigger. Interestingly, restoration of p16 into “fast” growing cells recovered the capacity to undergo TGFβ induced differentiation and express the typical differentiation markers. (Milyavsky et al., 2005).

    In addition to the loss of p16, we noticed a reduction in the myocardin gene. At that stage myocardin was shown to be a specific transcription factor that plays a central role in smooth muscle differentiation. Using several assays, we could show that the myocardin is lost at the point where slow growing cells are converted into fast growers we also found that myocardin expression is essential for their differentiation (Milyavsky et al., 2007). Furthermore, our studies have shown that myocardin over-expression caused cell growth arrest. Analysis of human cell lines and primary tumors indicated a trend of losing myocardin gene expression. We therefore concluded that at this early step of cell transformation, the cell had already undergone genetic changes that brought about a defect in their capacity to undergo cell differentiation. This defect is associated with a loss in the expression of at least two tumor suppressor genes, p16 and myocardin (Milyavsky et al., 2007; Shatz et al., 2007)

    Slow-growing cells

    The role of p53 in the process leading to such a cell differentiation blockage is presently being investigated in our laboratory. We expect that such an approach may indicate a defined role of p53 in cell differentiation at large.

    A third cluster of genes that we focused on consists of genes whose expression levels increased as a function of p53 and p16INK4A tumor suppressors' inactivation. This cluster predominantly consists of cell cycle-related genes and constitutes a signature of a diverse group of cancers. Promoters of the genes in this cluster are enriched with NFY, E2F, Elk1, CHR, and CDE regulatory motifs. The promoter architecture of many of the genes constitutes a “sum gate” – output mRNA levels correlate with the summated activity of the two suppressive channels of p53 and p16INK4A. Taking components of the mitotic spindle as an example, we experimentally verified our predictions that p53-mediated transcriptional repression of several of these novel targets is dependent on the activities of p21, NFY and E2F. Our study demonstrates how a well-controlled transformation process allows the linking three levels in a complex regulatory network, namely gene expression, promoter architecture and activity of upstream tumor suppressors (Tabach et al., 2005).

    Integrative view of gene expression pattern

    A fourth-interesting cluster observed in this study is a group of genes that are up-regulated by the over-expression of the RAS oncogene and concomitantly down-regulated by p53. This gene signature, which consists of a high number of chemokines, shows a significant increase and in some cases, a synergism in their expression as a result of overproducing RAS and are inactivated or knock-down of p53 expression. Based on the information obtained by this cluster expressed at an advanced phase of transformation we unraveled novel gene network that connects the RAS and p53.

    In general, it seems that the clusters that we have discovered using this in vitro model, seem to agree with specific stages in transformation and may thus serve as a specific hallmark signature of the stepwise malignant process.

    Gene signatures

  • We have recently also became interested in studying prostate cancer, which is the most commonly diagnosed type of cancer in men. There is yet no available cure for patients with advanced disease. Our working strategy is that developing an in vitro model will permit the development of new therapeutic modalities to address this issue. We have established, by immortalization, cultures of various prostate-derived cultures which were subjected to various malignant associated alterations. These cells exhibit a significant pattern of authentic prostate-specific features. In particular, the epithelial cell culture is able to differentiate into glandular buds that closely resemble the structures formed by primary prostate epithelial cells. The stromal cells have typical characteristics of prostate smooth muscle cells (Kogan et al., 2006). We have also immortalized prostate associated fibroblasts. These immortalized cultures may serve as a unique experimental platform to permit several research directions. This includs the study of cell-cell interactions in an authentic prostate micro-environment, prostate cell differentiation, and most significantly, the complex multi-step process leading to prostate cell transformation. A special focus is to resolve the role of p53 in the development of prostate, combined with a genome wide approach analysis.

    In summary, our in vitro systems offer interesting models that are expected to resolve and identify the various gene networks that are associated with defined steps of malignant transformation in general and their relevance to the p53 tumor suppressor gene. We accept that deciphering universal transcriptional programs, which are affected by the most common oncogenic mutations, will provide considerable insight into regulatory circuits controlling malignant transformation and will, hopefully, open new avenues for rational therapeutic decisions including p53-based therapy.