PerspectiveCancer

EGFR Signaling Inhibits E2F1-Induced Apoptosis in Vivo: Implications for Cancer Therapy

See allHide authors and affiliations

Science's STKE  30 Jan 2007:
Vol. 2007, Issue 371, pp. pe4
DOI: 10.1126/stke.3712007pe4

Abstract

The retinoblastoma tumor suppressor (RB) restricts cell proliferation by regulating members of the E2F family of transcription factors. In human tumors RB is often inactivated, resulting in aberrant E2F-dependent transcription and uncontrolled proliferation. One of the E2F proteins, E2F1, can also induce apoptosis. The extent of E2F1-induced apoptosis is known to be tissue- and cell-specific, but until now, it has been unclear what variables determine cellular sensitivity to E2F1-induced apoptosis in vivo. A recent study reveals epidermal growth factor receptor (EGFR) signaling to be one such variable, as EGFR signaling cooperates with RB in inhibiting E2F1-induced apoptosis. This finding raises the possibility that therapeutic manipulation of EGFR signaling may specifically trigger the death of cancer cells with inactive RB, thereby enabling "targeted" cancer treatments.

E2F transcription factors are best known for their ability to regulate the expression of genes required for DNA replication and cell cycle progression (1). However, it is now clear that E2Fs function in a wide range of biological processes, including differentiation, development, and DNA damage responses (2). In addition, at least one member of the E2F family, E2F1, can efficiently induce apoptosis both in cultured cells and in animal models (3). However, active E2F1 does not always induce cell death, and therefore the identification of physiological regulators of E2F1-induced apoptosis is a critical and current research challenge. A recent paper by Dyson and colleagues addresses this important issue and demonstrates that epidermal growth factor receptor (EGFR) signaling can inhibit E2F1-induced apoptosis in vivo (4). Furthermore, this study shows that the gradient of EGFR activity determines where and when during embryonic development RB-null cells undergo E2F1-dependent apoptosis.

E2F1 induces apoptosis through both p53-dependent and p53-independent pathways. Signaling from E2F1 to p53 also occurs through multiple mechanisms, but a primary signal transducer is ARF, a transcriptional target of E2F encoding a protein that stabilizes and activates p53 by interfering with the function of p53’s negative regulator, the E3 ubiquitin ligase Mdm2 (5). E2F1 induces phosphorylations of p53 that lead to p53 stabilization and activation (6). Furthermore, E2F1 up-regulates the expression of a number of proapoptotic cofactors of p53 and thereby directs p53 to induce apoptosis (7). The p53-independent apoptosis induced by E2F1 is attributed mainly to E2F-mediated activation of various proapoptotic genes, including those that encode Apaf-1 (apoptotic protease activating factor 1), caspases, BH3-only proteins, and the p53 family member p73 (3). Additionally, E2F1 sensitizes cells to apoptosis through inhibition of survival signals, in particular, those mediated by the transcription factor NF-κB or by Bcl-2 and its family member Mcl-1 (810). A conserved internal portion of E2F1 has a unique proapoptotic activity that distinguishes E2F1 from other E2Fs (11, 12).

E2F activity is regulated negatively by RB, the protein product of the retinoblastoma tumor suppressor gene. In most human tumors RB is functionally inactivated, resulting in deregulated and hyperactive E2F in the transformed cells. In this context, E2F1-induced apoptosis is viewed as a fail-safe mechanism that suppresses oncogenic transformation. In human tumors RB is often inactivated along with some part of the ARF-Mdm2-p53 pathway, and this may reflect selective pressures to reduce the otherwise apoptotic consequences of deregulated E2F1 activity.

However, loss of RB results in E2F1-dependent apoptosis only in some tissues and biological settings, and not in others. The variables determining whether deregulated E2F1 activity leads to apoptosis are not fully characterized. Because RB is often functionally inactive in human tumors, elucidating the conditions that favor selective apoptosis of cells with deregulated E2F1 would potentially advance cancer treatments. A number of in vitro studies show that serum attenuates E2F1-induced apoptosis in cultured cells, indicating that integration of external signals plays an important role in regulating E2F1-induced apoptosis. Furthermore, activation of the protein kinase AKT inhibits E2F1-induced apoptosis in mammalian cultured cells (12). However, until recently there was no evidence in vivo for such an effect of any signal transduction pathway on E2F1-induced apoptosis. The new study by Dyson and colleagues demonstrates that EGFR signaling regulates E2F1-induced apoptosis in vivo during the development of Drosophila imaginal eye discs (4).

The authors used the powerful tools of Drosophila genetics and took advantage of the simplicity of the RB-E2F families in the fly. Drosophila have one RB (rbf1) and two E2Fs (one activator, dE2F1, and one repressor, dE2F2), whereas humans possess three RB family members and eight E2Fs (three activators and five repressors). Detailed analysis of imaginal eye discs in rbf1-null flies revealed that cell death occurs at a specific time and location in the developing discs. Mutation of dE2F1 completely suppressed this cell death, indicating that the apoptosis resulting from rbf1 inactivation was E2F1-dependent. However, cell death did not occur in all cells exhibiting high E2F1 activity, which suggests that the deregulation of endogenous E2F1 was not sufficient to induce apoptosis. Dyson and colleagues found that down-regulation of the signaling pathway from EGFR to the small guanosine triphosphatase Ras and the protein kinase Raf was required for cell death. Cells with reduced activity of this pathway were highly sensitive to E2F1-induced apoptosis. Thus, this study shows that RB and EGFR-Ras-Raf signaling cooperate to inhibit E2F1-induced apoptosis during normal development. In rbf1-null cells, E2F1 activity is elevated but E2F1-induced apoptosis can be prevented by EGFR-Ras-Raf signaling.

How does the EGFR-Ras-Raf signaling pathway influence E2F1-induced apoptosis? A number of mechanisms can be envisaged. For example, the signal may impinge on E2F1 itself or on a critical proapoptotic target of E2F1, leading to their destruction or inhibition. Alternatively, the signal may act downstream of E2F1 targets by inhibiting their effects on the apoptotic process. In the Drosophila imaginal discs, the RB-E2F pathway and the EGFR-Ras-Raf pathway most probably converge on HID, a key regulator of cell death in Drosophila. HID is transcriptionally up-regulated by dE2F1 (13) and also inhibited by Ras and MAPK (mitogen-activated protein kinase)–induced phosphorylation (14). So one likely scenario is that loss of RBF1 results in increased abundance of HID, but HID activity is kept in check by EGFR-Ras-MAPK–mediated signaling. Once this signal is reduced, the high activity of HID causes apoptosis.

An important question concerns whether this functional link between the EGFR-Ras and RB-E2F pathways is evolutionarily conserved between flies and humans. Ras and E2F cooperate in transformation of mammalian cells, and expression of mutant forms of Ras and consequent activation of Raf inhibits apoptosis resulting from a triple knockout of all RB family members in mice (15), which suggests that this functional link is indeed conserved. However, it is not clear whether in humans it is mediated through the HID ortholog and whether it involves the MAPK pathway. Smac (also called Diablo), the best candidate for a human HID ortholog, is regulated transcriptionally by E2F in human cells (16), but it is not known whether Smac is regulated negatively by EGFR-induced phosphorylation. Another possible link between E2F and EGFR/Ras in humans is the proapoptotic BH3-only protein Bim, which is a transcriptional target of E2F (17) and is regulated negatively by a MAPK-induced phosphorylation (18). An effect of MAPK on E2F1-induced apoptosis remains to be demonstrated in mammalian cells.

The phosphatidylinositol 3-kinase–AKT pathway (which, like the Ras-MAPK pathway, is activated by the EGFR) inhibits E2F1-induced apoptosis in mammalian cells, but the MAPK pathway does not do so (12). Moreover, a recent study provides a possible link between AKT and E2F1. Lin and colleagues report that AKT phosphorylates TopBP1 (an adaptor protein named for its role as a DNA topoisomerase IIβ binding protein) and that this phosphorylation of TopBP1 allows TopBP1 to interact with E2F1 and inhibit its apoptotic activity (19). So, although an effect of Ras to inhibit E2F1-induced apoptosis seems to be evolutionarily conserved, the mechanism of signal cross-talk may differ between flies and humans (Fig. 1). E2F1 activates AKT in human cells, suggesting the existence of an AKT-E2F1 feedback loop by which E2F1 might modulate its own apoptotic activity (20). It remains to be determined whether similar feedback loops function in vivo.

Fig. 1.

Regulation of E2F1-induced apoptosis in (A) Drosophila and (B) humans. In both species, E2F1-induced apoptosis is inhibited by a cooperation between RB and signal transduction pathways. However, the molecular mechanism of the cross-talk between these signaling pathways may be tissue- and species-specific.

The discovery that the EGFR-Ras pathway inhibits E2F1-induced apoptosis in flies has potentially important implications for future cancer therapy. Pharmaceutical attenuation of signaling pathways that inhibit E2F1-induced apoptosis may lead to selective killing of tumor cells that lack functional RB but retain operational apoptotic machinery downstream of E2F1. However, several challenges lie ahead. The distinct signaling pathways that regulate E2F1-induced apoptosis in specific human tissues and tumors must be characterized to enable proper design of tissue culture–based screens for such proapoptotic drugs. Furthermore, delineation of the molecular mechanisms underlying the cross-talk between signaling pathways and E2F1 activity is required, as this will likely reveal key therapeutic targets.

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.
View Abstract

Navigate This Article