PerspectiveGene Silencing

Silent Assassin: Oncogenic Ras Directs Epigenetic Inactivation of Target Genes

See allHide authors and affiliations

Science Signaling  01 Apr 2008:
Vol. 1, Issue 13, pp. pe14
DOI: 10.1126/stke.113pe14

Abstract

Oncogenic transformation is associated with genetic changes and epigenetic alterations. A study now shows that oncogenic Ras uses a complex and elaborate epigenetic silencing program to specifically repress the expression of multiple unrelated cancer-suppressing genes through a common pathway. These results suggest that cancer-related epigenetic modifications may arise through a specific and instructive mechanism and that genetic changes and epigenetic alterations are intimately connected and contribute to tumorigenesis cooperatively.

Neoplastic transformation arises from the accumulation of genetic and epigenetic alterations. These changes result in the activation of oncogenes and inactivation of tumor suppressor genes. Collaboration among various oncogenes and the loss of tumor suppressors disrupts cellular homeostasis and induces the progressive transformation of a normal cell into its evil twin, a malignant cancer cell (1). Although it is reasonable to assume that genetic and epigenetic changes are integrated and interdependent processes during tumorigenesis, the mechanistic link between genetic changes and epigenetic alterations is largely unclear. An elegant study by Gazin et al. fills the gap in our understanding in this specific area (2).

DNA methylation at the C-5 position of cytosine in CpG dinucleotides, which is catalyzed by DNA methyltransferases (DNMTs), is the most common and important epigenetic alteration. DNA methylation is critical for transcriptional regulation, modulation of chromatin structure, and genomic imprinting (3). In cancer cells, DNA methylation patterns are substantially altered, with an overall hypomethylated genome and hypermethylation of CpG islands at localized DNA regions (47). Whereas global hypomethylation contributes to carcinogenesis through activation of oncogenes (810), loss of imprinting (11, 12), and chromosome instability (1315), epigenetic silencing of tumor suppressor genes and other cancer-related genes by hypermethylation is the most frequent molecular lesion in cancers. However, it is not clear whether specific epigenetic signatures in cancer are the result of the selection of randomly acquired epigenetic changes or are due to targeted modifications that are controlled by specific pathways. To gain further insight into these processes, Gazin and colleagues used a genomewide RNA interference (RNAi) screen to systematically analyze oncogenic Ras-mediated epigenetic regulation (2).

The small guanosine triphosphatase (GTPase) Ras is perhaps the best-known oncogene, and it plays a critical role in tumor growth and development. More than 30% of all human cancers contain activating Ras mutations (16). Activation of this pathway in the absence of Ras mutations also frequently occurs in various human tumors because of alterations in signaling components upstream or downstream of Ras (17). In addition to promoting abnormal cell proliferation, oncogenic Ras epigenetically represses the expression of an array of cancer-related genes involved in apoptosis, cell cycle regulation, DNA repair, and differentiation. These include the genes that encode Fas (18, 19), p16 tumor suppressor (20), lysyl oxidase (21), transforming growth factor β (TGF-β) receptor type II (10), prostate apoptosis response 4 (Par4) (22), the cytoskeletal protein tropomyosin (23), and the tumor suppressor protein opioid binding protein/cell adhesion molecule–like gene (OPCML) (24). However, the mechanism of Ras-mediated epigenetic silencing is not well understood. Taking advantage of the cell surface localization of Fas, Gazin and colleagues designed a clever genomewide short hairpin RNA (shRNA) screening strategy. They transduced K-ras–transformed NIH 3T3 cells, which do not express Fas at the cell surface, with retroviruses that contained shRNAs and subsequently selected those cells that regained surface expression of Fas in response to shRNA-mediated gene silencing with immunomagnetic beads coated with a specific antibody against Fas. To streamline the screening process, the authors used 10 pools of retroviral particles, each containing about 6000 mouse shRNA clones, to individually silence the expression of 28,000 genes in the K-ras–transformed NIH 3T3 cells. The authors identified candidate genes that were important for Ras-mediated Fas silencing by analyzing the sequences of amplified shRNA fragments derived from the bead-selected cells in each pool. To confirm their screening results, Gazin et al. verified the candidate genes individually by using single gene-specific shRNAs. The authors eventually identified 28 genes whose silencing by shRNA led to the re-expression of Fas at the cell surface. To eliminate potential complications resulting from off-target and cell-type specific effects, the authors performed the same experiments with an unrelated set of shRNAs that targeted each of the 28 candidate genes, as well as in an unrelated cell line, H-ras–transformed mouse C3H10T1/2 cells. These results convincingly demonstrated the existence of a group of Ras epigenetic silencing effectors (RESEs) that were responsible for Ras-mediated epigenetic silencing of Fas (2).

Although the identification of several well-known molecules associated with Ras signaling pathways among the identified RESEs was reassuring, the involvement of a large group of nuclear gene regulatory proteins in Ras-mediated Fas silencing was a pleasant surprise. These nuclear RESEs included transcriptional regulators and DNA-binding proteins, proteins involved in histone and DNA modifications, and several Polycomb group proteins. Gazin and colleagues showed that several RESEs, including DNMT1, were up-regulated at the transcriptional or posttranscriptional level in K-ras–transformed NIH 3T3 cells, providing a potential mechanism to explain how oncogenic Ras might activate the silencing pathway (2). These findings are consistent with elevated DNMT1 activity being important for Ras-mediated oncogenic transformation (25, 26). Interestingly, the majority of nuclear RESEs represented novel components of the Ras signaling pathway (2).

To verify the involvement of DNMT1 or DNA hypermethylation in Ras-mediated Fas silencing, the authors showed that the Fas promoter was indeed hypermethylated at three specific regions in K-ras–transformed NIH 3T3 and H-ras–transformed C3H10T1/2 cells. On the other hand, the same regions were largely unmethylated in nontransformed NIH 3T3 and C3H10T1/2 cells and in K-ras–transformed NIH 3T3 and H-ras–transformed C3H10T1/2 cells in which DNMT1 had been knocked down by shRNA. Treatment of K-ras–transformed NIH 3T3 cells with the DNMT inhibitor, 5-aza-2′-deoxycytidine (5-aza) or with shRNAs against individual RESEs led to the demethylation of the Fas promoter and the re-expression of Fas at the cell surface (2).

How does oncogenic Ras specifically program the epigenetic silencing of Fas? One potential mechanism is through the active recruitment of the gene-silencing machinery to the Fas promoter, as has been demonstrated in PML-RAR– [a fusion protein of promyelocytic leukemia (PML) and the retinoic acid receptor (RAR)] and Myc–mediated transcriptional repression of RARβ2 and p21, respectively (27, 28). Gazin et al. systematically carried out chromatin immunoprecipitation (ChIP) assays with three sets of primer pairs covering the entire 2-kB Fas promoter to determine whether nuclear RESEs bound directly to the Fas promoter. They found a total of nine Fas promoter–associated RESEs in K-ras–transformed NIH 3T3 cells, whereas only one of the nine RESEs, nucleophosmin/nucleoplasmin 2 (NPM2), was associated with the Fas promoter in untransformed NIH 3T3 cells. DNMT1 was among the nine RESEs that bound to the Fas promoter, and this association was dramatically reduced in each of 28 K-ras–transformed NIH 3T3 cell lines in which individual RESEs had been knocked down by shRNA. Conversely, the other two de novo DNMTs, DNMT3a and DNMT3b, were not noticeably associated with the Fas promoter as determined by ChIP analyses. These results suggested that DNMT1 is a key molecule required for maintaining Ras-mediated silencing of Fas and promoter hypermethylation (2). It is particularly interesting that several Polycomb group proteins, including the histone methyltransferase (EZH2), were among the identified RESEs. Other studies show that EZH2-containing Polycomb complexes act as recruitment platforms for DNMTs and direct specific DNA methylation by marking the targeted promoters with methylated histone H3 (29, 30). Therefore, EZH2 may serve as a potential downstream effector through which Ras mediates target-specific epigenetic silencing.

To ensure that the functions of RESEs were not just limited to Fas, the authors performed parallel analyses of a set of five well-documented, epigenetically silenced genes: Sfrp1 (secreted frizzled-related protein 1), Par4, Plagl1 (pleiomorphic adenoma gene–like 1), H2-K1 (histocompatibility 2, K1, K region), and Lox (lipoxygenase) that contribute to oncogenic transformation and cancer. The expression patterns of these five genes and the methylation status of their promoters were similar to those of Fas found in the earlier experiments. Most importantly, Gazin et al. demonstrated that, of the 28 RESEs required for silencing Fas, at least 21 were also essential for promoter hypermethylation and silencing of each of these five other target genes. These results suggested that Ras directed the epigenetic silencing program for the five target genes with a common cast of characters and a similar script to that used for Fas silencing. Lastly, the authors showed that some of the 28 RESEs were also important for Ras-mediated oncogenic transformation. Knocking down these RESEs individually inhibited the ability of K-ras–transformed NIH 3T3 cells to grow on soft agar and to form tumors in nude mice (2).

The study by Gazin et al. revealed that oncogenic Ras used a specific and complex epigenetic silencing program that involved an elaborate network of partners with diverse functions. These findings provide a direct mechanistic link between oncogene activation and epigenetic silencing, supporting the notion that seemingly chaotic epigenetic silencing events are programmed by specific and well-defined pathways during tumorigenesis. Moreover, demonstration of pathway-specific Ras-mediated epigenetic silencing suggests that oncogene activation (genetic changes) and epigenetic alterations are well-integrated and cooperative processes in pathways to cancers (Fig. 1).

Fig. 1.

Mechanism of oncogenic transformation. (A) Cooperation between genetic changes and epigenetic alterations leads to the activation of oncogenes and the inactivation of tumor suppressors, which results in tumorigenesis. (B) Activation of oncogenic Ras (RasV12) directs a specific epigenetic silencing program during tumorigenesis by recruiting epigenetic modification enzymes, such as DNMT1, and hypermethylating specific promoters of cancer-related genes.

Like other important scientific discoveries, the study by Gazin and colleagues also raises many questions. First, can other oncogenes direct similar epigenetic silencing programs like Ras? Although additional studies are required to answer this question, the existence of common epigenetic silencing targets in cells transformed by different oncogenes suggests that general oncogene-mediated epigenetic mechanisms may exist during tumorigenesis (26). Second, can Ras initiate the epigenetic silencing program independently, or are other accomplices required? It is well known that in some primary cells oncogenic Ras triggers cellular senescence, instead of cellular transformation, by inducing the accumulation of p53 and p16 (31). Therefore, it is very likely that activation of the Ras-mediated epigenetic silencing program may require additional inputs or a specific cellular setting. Third, does Ras use the identical epigenetic silencing program in human cancers? Gazin and colleagues performed all of their studies with murine cell lines. Although humans and mice share most if not all of the components and signaling pathways described, important differences in the requirements for neoplastic transformation exist between humans and rodents (32). A careful evaluation of the findings by Gazin et al. in human systems is necessary before extrapolating their findings to human cancers. Lastly, what are the roles that histone modifications, particularly histone deacetylation, play in Ras-mediated epigenetic silencing? Whereas Gazin and colleagues explored the role of DNMT1 and DNA methylation extensively in their study, they did not investigate the involvement of histone deacetylases (HDACs) and histone modification. Considering that several histone-modification enzymes were among the identified RESEs and that epigenetic silencing requires collaboration between histone deacetylation and DNA methylation, the answer to this specific question will be important for further understanding the mechanism of Ras-mediated epigenetic silencing. Despite these questions, the study by Gazin and colleagues has substantially extended our knowledge of oncogene activation and epigenetic regulation in tumorigenesis. The identification of novel effectors downstream of Ras that are important for epigenetic silencing and tumorigenesis may one day lead to the development of new anticancer therapies that specifically target the Ras pathway, a difficult task that has evaded success for many years.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
View Abstract

Stay Connected to Science Signaling

Navigate This Article