Preuzmite MALIGNA CELIJA-skripta-predavanje Part3 i više Rezime u PDF od Biologija ćelije samo na Docsity!
Groudine, reported that the c-myc gene was amplified about tenfold in the human promyelocytic leukemia cell line HL-60. Amplification of c-myc was detected not only in the tumor cells after growth in culture, but also in the primary uncultured leukemic cells from which the HL-60 cell line was derived. This amplification of c-myc DNA was associated with high levels of c-myc transcription in HL-60 cells. In addition, cytological studies and hybridization of c-myc probe to preparations of metaphase chromosomes indicated that the amplified c-myc sequences in HL-60 cells were associated with either double minutes or abnormal chromosome staining regions typical of DNA amplification. Amplification and overexpression of c-myc are not restricted to HL-60 cells but have also been detected in a number of other tumors including breast, stomach, and lung carcinomas, malignant neuroendocrine cells derived from a colon carcinoma, neuroblastomas, and glioblastomas (Table 8.1). Amplification of c-myc is particularly common in breast cancers and small-cell lung carcinomas, occurring with frequencies of about 30%. Because c-myc has been encountered as an activated oncogene in a wide variety of different settings, its elevated expression as a correlate of gene amplification in a number of human neoplasms provides strong support for a role of DNA amplification in the activation of cellular oncogenes in the course of tumor development.
N-myc amplification and neuroblastoma prognosis The N-myc gene is often amplified in human childhood neuroblastomas. Multiple copies of this gene has been detected (yellow) by FISH. The fact that these copies are present as tandem arrays (yellow) means that they constitute HSR rather than DMs which are also frequently seen in these tumors
Myc protein is a transcription factor that activates expression of a great number of genes through binding on consensus sequences (Enhancer Box sequences (E-boxes)) and recruiting histone acetyltransferases (HATs). It can also act as a transcriptional repressor. By binding Miz-1 transcription factor and displacing the p300 coactivator, it inhibits expression of Miz-1 target genes.
Oncogenes activated by locus amplification
- Amplified sequences can be seen in karyotypes as:
- double minute (DM) chromosomes - very small accessory chromosomes
- additional banding regions called homogeneously staining regions (HSR)
- 20-100s of copies of a DNA region of several hundred thousand bases-extra copies of proto-oncogenes - NMYC, HER
The plot illustrates the event-fee survival (EFS) on children suffering neuroblastoma, i.e. no clinically significant cancer-related observation or occurrences in the indicated years following initial diagnosis or treatment. Thos who have minimal or no N-myc amplification have a very good prognosis while those with extensive amplification have dramatically poorer prognosis and therefore short survival time after diagnosis.
New Members of the myc Gene Family DNA amplification, as evidenced cytologically by double minute chromosomes and homogeneous staining regions, is a particularly common feature of neuroblastomas. Consequently, this neoplasm appeared to serve as an especially attractive model for studies directed toward understanding the potential significance of amplified DNA to tumor development. In 1983, two research groups reported that the DNA sequences amplified in many neuroblastomas cross-hybridized with a c-myc probe. The amplified DNA, however, was distinct from the c-myc gene and was only readily detected with c-myc probe when enhanced hybridization resulted from the high number of amplified copies present in neuroblastoma DNAs. Because of its relationship to c-myc and its reproducible amplification in neuroblastomas, the sequence identified in these studies appeared to represent a new candidate oncogene, which was designated N-myc. As expected, the N-myc sequences were localized to double minutes and homogeneous staining regions and were transcribed at high levels in those neuroblastomas in which the gene was amplified. Subsequent molecular cloning and characterization has verified that N-myc encodes a protein that is related to but distinct from the c-myc gene product (Fig. 8.2). In addition, gene transfer experiments have demonstrated the ability of N-myc to function in the induction of cell transformation. The N-myc gene is thus a new member of the myc oncogene family. Its amplification in neuroblastomas is not only a frequent event, but also highly correlated with the progression of these tumors to increasingly malignant stages associated with a poor clinical prognosis. These studies have provided some of the strongest available evidence for a relation between cellular oncogene activation and the clinical behavior of a human malignancy. Amplification and expression of N-myc has also been detected in other neoplasms, including retinoblastomas and small-cell lung carcinomas (which also appear to be neoplasms of neuroendocrine origin) (Table 8.1). As noted earlier, amplification of c-myc is also a frequent event in lung cancers, particularly small-cell carcinomas. In addition, studies of these neoplasms revealed amplification of a third member of the myc gene family
This contrasts with amplification of myc genes, because it had been directly demonstrated by gene transfer experiments that constitutive expression of a normal myc protein is sufficient to elicit cell transformation (see chapter 7). The consequence of overexpression of the normal erbB proto-oncogene was therefore investigated by gene transfer experiments, using a vector that led to expression of high levels of the normal erbB gene product (the epidermal growth factor receptor) in transfected cells. Analysis of these transfected cells indicated that their normal proliferative response to epidermal growth factor was enhanced and that, in the presence of epidermal growth factor, they displayed a transformed phenotype. These results thus provided direct support for the significance of amplification and overexpression of the normal erbB proto-oncogene in neoplastic transformation. The human gene designated erbB-2 was initially isolated in 1985 by cross- hybridization with erbB probe. As was the case in studies that led to the isolation of the N- myc and L-myc genes, hybridization of erbB probe to DNA of a breast carcinoma cell line revealed amplification of a cross-hybridizing gene that was distinct from the cellular erbB oncogene. Amplification of the same erbB-related sequence in a human salivary gland carcinoma was independently reported at the same time by a second group of researchers. Molecular cloning and sequence analysis confirmed that the new gene, designated erbB-2, was related to but distinct from erbB, apparently encoding a cell surface receptor for a distinct ligand (Fig. 8.3). INTERESTINGLY, THIS ANALYSIS ALSO REVEALED THAT erbB- 2 WAS THE HUMAN HOMOLOG OF THE GENE ORIGINALLY CALLED NEU, WHICH HAD FIRST BEEN DETECTED BY GENE TRANSFER EXPERIMENTS AS AN ACTIVATED ONCOGENE IN RAT NEUROBLASTOMAS (see chapter 5).
Activation of the transforming potential of erbB-2 (neu) in rat neuroblastomas is a consequence of a point mutation that yields a structurally altered gene product. As with erbB, it was therefore important to investigate the biological significance of amplification and overexpression of the normal erbB-2 protein. Gene transfer experiments with appropriate molecular constructs have, however, demonstrated that high-level expression of normal erbB-2 is indeed sufficient to induce cell transformation.
Amplification of erbB-2 is particularly common in human breast and ovarian cancers, having been detected in -25% of these neoplasms. As further discussed in chapter 11, erbB-2 amplification is closely correlated with the clinical behavior of these neoplasms, such that tumors with erbB-2 amplification are more aggressive and associated with reduced patient survival.
Other Amplified Oncogenes In addition to amplification of c-myc and erbB-2, amplification of three other oncogenes has been detected in about 20% of human breast carcinomas, as well as in several types of squamous cell carcinomas.
Amplification of the erbB2/HER2/neu oncogene in breast cancers HER2 is thought to be an orpfan receptor HER2 is thought to be an, with none of the EGF family of ligands able to activate it. However, ErbB receptors dimerise on ligand binding, and HER2 is the preferential dimerisation partner of other members of the ErbB family. , with none of the EGF family of ligands able to activate it. However, ErbB receptors dimerise on ligand binding, and HER2 is the preferential dimerisation partner of other members of the ErbB family.
These three genes ( list, int-2, and PRAD-1 ), which had previously been identified as oncogenes by other approaches, are all closely linked on chromosome 11 and are frequently amplified as a unit. The amplification of int-2 in human breast cancers is particularly interesting because int-2 was first identified as a cellular oncogene activated by promoter insertion in mouse mammary carcinomas (see chapter 6). However, neither int- nor hst is generally expressed in those tumors in which it is amplified, so it appears unlikely that either is the functional target of the chromosome 11 amplification. PRAD-1, on the other hand, is overexpressed in those breast and squamous cell carcinomas in which it has been amplified, suggesting that it is the key oncogene within this amplified unit of DNA. A number of other known cellular oncogenes also have been found to be amplified, but amplification of these genes occurs only in occasional tumors or tumor cell lines, suggesting that they do not play a reproducible role in tumorigenesis. These genes include all three members of the ras gene family (rasH, rasK, and rasN), abl, ets-1, and myb. In some cases, the amplified oncogenes have also sustained additional lesions that may activate their transforming potential. For example, the chronic myelogenous leukemia cell
CANCER AND TUMOR SUPPRESSOR GENES FOCUS ON RB TS
Function – restrain cell proliferation Cancer - Absence or Inactivation of tumor suppressor genes
Tumor Suppressor Genes
Rb Gene - Gene discovered in heriditary retinoblastoma
- chr. 13
- One mutated allele inherited
- Other allele mutates or is deleted during cell division
- In sporadic disease the two Rb alleles are separately inactivated as two separate genomic events
Retinoblastoma and the Discovery of Tumor Suppressor Genes Cellular oncogenes were ultimately defined by their ability to induce neoplastic transformation upon introduction into an appropriate target cell. These genes act in a dominant positive regulatory fashion in that the introduction of a functional oncogene into a non-transformed recipient cell results in development of a new phenotype-transformation in the presence of TWO COPIES OF THE NORMAL ALLELES. Moreover, oncogene activation is a consequence of mutations or gene rearrangements that increase gene expression or increase the activity of the oncogene encoded proteins. In contrast with the dominant oncogenes, a distinct class of genes that function as negative regulators of neoplastic disease has been revealed by studies of somatic cell hybrids and of some inherited human cancers. For this class of genes, loss of function leads to tumor development. Although molecular analysis of these genes is less advanced than
studies of the dominant oncogenes, it is clear that loss of function of a number of cellular genes plays a critical role in the pathogenesis of many human neoplasms. Such genes have been referred to as cancer susceptibility genes, anti-oncogenes, recessive cancer genes, recessive oncogenes, and (as used here) tumor suppressor genes.
Somatic Cell Hybridization The first approach to genetic analysis of neoplastic cells, initiated in 1969 by Henry Harris and his colleagues, was to fuse normal cells with tumor cells and analyze the properties of the resulting hybrids. Extensive studies of this type have now established that most hybrids between normal and malignant cells are no longer tumorigenic. Such suppression of tumorigenicity by fusion of a tumor cell with a normal cell implies that normal cells contain one or more genes that act as negative regulators of the neoplastic phenotype: the tumor suppressor genes. Interestingly, however, many nonturnorigenic hybrids between normal and malignant cells still display some phenotypic properties of transformed cells in vitro, including loss of density-dependent inhibition of cell growth and loss of anchorage dependence. Thus, it appears that suppression of the tumorigenic phenotype in hybrids constitutes only partial, rather than complete, reversion of cell transformation. As discussed later, this is consistent with the notion that both activation of dominant oncogenes and loss of tumor suppressor gene function independently and contribute to development of the fully malignant phenotype of many tumors. The interpretation that the nontumorigenic phenotype of hybrids between normal and malignant cells results from the action of specific normal cell genes is strengthened by the fact that such hybrids frequently revert to the tumorigenic phenotype following loss of specific chromosomes of the normal parent. For example, hybrids between normal human fibroblasts and HeLa cells (a human cervical carcinoma cell line) are nontumorigenic but revert to the tumorigenic phenotype following loss of normal chromosome 11. These findings suggest that tumorigenicity of the malignant parent (HeLa cells) resulted, at least in part, from loss of function of a tumor suppressor gene located on chromosome 11. Introduction of the wild-type allele by fusion with a normal cell would then restore function of the gene and suppress tumorigenicity. If the relevant normal chromosome 11 was subsequently lost, the hybrid would revert to the tumorigenic phenotype. Similar studies of hybrids between normal cells and other tumor cell lines have identified additional human chromosomes that carry tumor suppressor genes that may have been inactivated in other human neoplasms. Somatic cell hybridization experiments have thus provided a clear indication that tumorigenicity is associated with loss of function of critical regulatory genes in malignant cells. It remained, however, for such tumor suppressor genes to be defined at the molecular level by a different approach - namely, the analysis of inherited human cancers.
sporadic form of the disease are generally older when they develop tumors than children with inherited disease. Although susceptibility to retinoblastoma is transmitted in a genetically dominant manner, inheritance of the susceptibility gene does not by itself suffice to transform a normal retinal cell into a tumor cell. Thus, patients with inherited retinoblastoma develop only a few focal tumor growths over a background of several million normal retinal cells. Even though all retinal cells have inherited the susceptibility gene, most are unaffected and function normally: only a small fraction actually become neoplastic thus DEMONSTRATING RECESSIVE NATURE OF THE GENE.
Multi-hit Cancer theory: Also in hereditary , One hit inherited , One hit mutation
This pedigree shows multiple generations of a kindred afflicted with familial retinoblastoma, a disease that usually strikes only 1 in 20,000 children. Such multiple-
generation pedigrees were rarely observed before the advent of modern medicine, which allows an affected child to be cured of the disease and therefore reach reproductive age. Males (squares), females (circles), affected individuals (green filled circles, squares), unaffected individuals (open circles, squares).
Nonheritable vs Heritable Retinoblastoma
Feature Nonheritable Heritable
Tumor Unilateral Usually bilateral
Family history None 20% of cases
Increased risk of second primaries None
Osteosarcoma, other sarcomas, melanoma, others
It therefore appears that, at the cellular level, inheritance of the retinoblastoma susceptibility gene is not sufficient for neoplastic transformation and that additional alterations must occur before the tumor phenotype is expressed. Based on statistical analysis of the comparative frequency and age of development of inherited and sporadic forms of retinoblastoma, Alfred Knudson in 1971 proposed that retinoblastoma is caused by two mutations. In inherited disease, the first mutation is present in the germ line and the second mutation occurs somatically, resulting in tumor development. The probability of the second mutation is sufficiently high that tumors will nearly always develop among the more than 106 cells in each retina. In sporadic disease, both required mutations are somatic and only a very rare retinal cell in which two independent mutations have occurred will become neoplastic. This two mutation model accounted for the dominant inheritance of susceptibility to retinoblastoma in spite of the fact that the susceptibility gene did not function as a single dominant determinant of transformation at the cellular level. However, the model did not address the nature of the gene or genes involved. One possibility that would be compatible with Knudson's two-step model is that the first mutation activated a dominant oncogene and that a second mutation, perhaps activating a second oncogene, was required for development of neoplasms. Multiple events in tumor progression were alluded to earlier and several model systems (for example, transgenic mice carrying activated oncogenes) in which oncogene activation appears to be one step in a multistep pathway of tumor development will be discussed. However, this is not the case for retinoblastoma. Rather, the two mutations required for retinoblastoma development have proved to be the loss of both functional copies of the retinoblastoma susceptibility gene that would be present on complementary chromosomes in a normal diploid cell. These studies have thus revealed that loss rather than activation of
Karyotypic analysis also identified a locus on chromosome 13ql4 as being involved in the pathogenesis of SPORADIC retinoblastoma. Chromosome 13 deletions involving band 13ql4 were found in the tumors of about 25% of patients with the noninherited form of the disease. In these cases, normal blood cells of the same patients contained TWO NORMAL chromosome 13 homologs, INDICATING that deletion of band 13ql4 had OCCURRED AS A SOMATIC MUTATION in the course of tumor development. Thus, deletions of the same chromosomal locus appeared to be involved in at least some cases of both inherited and sporadic retinoblastoma. Although deletions of chromosome 13ql4 seemed to consistently occur in some cases of retinoblastoma, detectable karyotypic abnormalities were not found in the majority of cases. However, a deletion large enough to be recognized at the level of chromosome morphology would have to extend over hundreds of kilobases of DNA. Clearly, loss of gene function could occur as a consequence of smaller deletions (or point mutations) that would not be detectable by karyotypic analysis. In order to investigate the potential contribution of such mutations to the majority of retinoblastomas, which did not evidence gross chromosomal deletions, a fine structure genetic approach to defining the retinoblastoma locus was needed. This was accomplished by analyzing the linkage of the retinoblastoma susceptibility gene to another genetic marker, the gene encoding esterase D, which had independently been mapped to the same chromosomal locus. Two different alleles of this gene could be distinguished because they encoded electrophoretically separable forms (isozymes) of esterase D. Thus, heterozygosity at the esterase D locus provided a linked marker gene that could be followed in families with inherited retinoblastoma. Such studies demonstrated close linkage between esterase D and retinoblastoma development in families without recognizable chromosome deletions (Fig. 9.5). The children in these families who developed retinoblastoma ALWAYS inherited the gene for one esterase D isozyme: children who did not develop retinoblastoma inherited the other. These findings indicated that susceptibility to retinoblastoma was inherited through a gene on chromosome 13, even in cases without detectable chromosome deletions.
The Retinoblastoma Gene is Recessive at the Cellular Level The mapping of the retinoblastoma susceptibility gene to a chromosomal locus, which was DELETED IN SOME CASES of BOTH INHERITED AND SPORADIC DISEASE, SUGGESTED THE POSSIBILITY that the critical event in development of this neoplasm was LOSS OF A FUNCTIONAL GENE. If this were the case, the two mutations required for retinoblastoma development might represent inactivation of both copies of the wild-type gene, present on two chromosomes of a normal cell. The first mutation - whether germ line or somatic - would result in a cell with one mutant and one normal allele. The second mutation would inactivate the remaining normal copy of the gene, leading to complete loss of gene function. Inactivation of the second wild-type allele in a cell already containing one mutant allele could occur either by an independent somatic mutation or by loss of the chromosome region containing the remaining normal locus (Fig. 9.6). The latter class of events, which could occur either by mitotic recombination or by chromosome loss with or without duplication of the remaining chromosome, would result in tumors that would be homozygous for the retinoblastoma gene and for other linked markers on chromosome 13. Two different experimental approaches VERIFIED THIS PREDICTED HOMOZYGOSITY in tumors and thus provided direct evidence that neoplasia results from loss of retinoblastoma gene function.
The first approach again used esterase D as a linked genetic marker by comparing esterase D genes in normal somatic cells with retinoblastomas of the same patients (Fig. 9.7). Such studies identified a number of cases in which NORMAL CELLS WERE HETEROZYGOUS FOR ESTERASE D, WHEREAS ONLY ONE OF THE TWO INHERITED ALLELES WAS PRESENT IN THE NEOPLASMS. Thus, retinoblastomas were frequently REDUCED TO HOMOZYGOSITY AT THE ESTERASE D LOCUS
hybridizes to fragments of varying sizes in DNAs of different individuals identifies an RFLP, which provides a polymorphic genetic marker defined by both a probe and a particular restriction endonuclease. Such markers are more easily identified because they are not dependent on finding functional differences between expressed proteins-for example, isozymes of esterase D. Consequently, a large number of RFLPs have now been mapped to specific regions of human chromosomes and are used as markers in genetic linkage analysis. The application of RFLPs to analysis of retinoblastomas was first reported by Webster Covance, Ray White, and their collaborators in 1983 (Fig. 9.8). These researchers used RFLPs derived from chromosome 13 to compare the DNAs of normal blood cells and retinoblastomas from a series of patients. In several cases, they identified individuals who were heterozygous for an RFLP in their blood cells but homozygous for the same RFLP in their tumors. It was also found in some cases of inherited retinoblastoma that the copy of chromosome 13 from the affected parent (and thus bearing the retinoblastoma mutation) was always the copy retained in the tumors.
A further step in this analysis came from studies of Thaddeus Dryja and his colleagues in 1986. By screening retinoblastomas with additional probes derived from the 13ql4 region, they identified tumors with homozygous deletions of at least 25 kb of DNA. Neoplasm development was therefore unambiguously associated with deletion and complete loss of retinoblastoma gene function. Taken together, these genetic studies clearly established that retinoblastoma development was due to the INACTIVATION OR LOSS OF BOTH COPIES OF THE RETINOBLASTOMA SUSCEPTIBILITY GENE (gen predispozicije za retinoblastom). THESE EXPERIMENTS THUS DEFINE A NEW CLASS OF GENES WHOSE LOSS OF FUNCTION (RATHER THAN WHOSE ACTIVATION) CAN LEAD TO NEOPLASTIC TRANSFORMATION.
Molecular Cloning of the Rb Gene Although the foregoing genetic experiments elegantly defined a gene for retinoblastoma susceptibility, the application of genetics alone could not elucidate the nature or function of the gene product. However, the chromosomal localization of the retinoblastoma gene (Rb), and the finding of homozygous deletions in some tumors, provided a plausible approach to the isolation of the gene as a molecular clone, a step that would begin to extend the genetic studies to the molecular and biochemical levels of understanding. Cloning the Rb gene was undertaken by three independent groups - a collaboration between the laboratories of Thaddeus Dryja and Robert Weinberg; the laboratory of WenHwa Lee; and the laboratory of Yuen-Kai Fung and William Benedict - who reported isolation of the gene in late 1986 and early 1987. The general strategy used by all of the research groups was the same: existing probes to the 13ql4 region were used to isolate a series of genomic clones of the surrounding DNA. These clones were then used as probes for hybridization to RNAs of normal retinal cells and retinoblastomas. Each group identified candidate probes that hybridized to a normal retinal transcript of -4.7 kb that was not detected in RNAs of at least some retinoblastomas. The expression of this 4.7 kb RNA in normal retinal cells (and a variety of other cell types) but its absence in some retinoblastomas indicated that it was a plausible candidate for the transcript of the retinoblastoma gene. Consequently, cDNA clones of the candidate transcript were isolated and used for further characterization of gene structure and expression. The structure was studied by screening of genomic libraries and the expression by the analysis of RNA and protein. These studies detected frequent deletions of the candidate gene in retinoblastomas. Importantly, in some cases, the boundaries of the deletions were contained entirely within the candidate gene—these internal deletions strongly suggested that the critical target was the cloned gene itself rather than a flanking gene. In addition, many retinoblastomas that did not display deletions of the candidate Rb gene were nevertheless abnormal in its transcription, either lacking detectable levels of RNA or expressing abnormally sized transcripts. Inactivation of the gene at the level of transcription could therefore be detected in tumors without gross deletions, presumably as a consequence of small deletions or point mutations. Finally, point mutations were detected within the candidate gene even in those tumors expressing normal levels of the 4.7 kb RNA.
Role of the Rb Gene in Other Neoplasms Although the Rb gene was initially identified by inherited susceptibility to retinoblastoma, the loss of function of this gene also contributes to the development of several other tumors. Surviving patients with inherited retinoblastoma develop second neoplasms with much higher frequencies than either normal individuals or survivors of sporadic retinoblastoma. Nearly half of the secondary neoplasms in inherited retinoblastoma patients are osteosarcomas, suggesting that loss of Rb function also predisposes to osteosarcoma development. Molecular analyses have supported this notion. Both inherited and sporadic osteosarcomas are frequently homozygous for chromosome 13 RFLPs, parallel to the situation in retinoblastoma. In addition, the Rb gene itself is deleted or otherwise mutated in a high fraction of osteosarcomas. Thus, a single tumor suppressor gene predisposes to the development of both of these inherited neoplasms. Inherited Rb mutations are rare and retinoblastoma occurs with a low frequency, affecting only about 1 in 20,000 children. However, mutations of the Rb tumor suppressor gene also contribute to some of the common malignancies of adults, including lung, breast, and bladder cancers. In these tumors, which are more than 1,000 times as frequent as retinoblastoma, the inactivating mutations of Rb occur somatically rather than being inherited. In fact, these tumors are not common among survivors of inherited retinoblastoma, suggesting that mutations of genes other than Rb limit their development. However, because Rb mutations are found in a significant fraction of these malignancies, the role of the Rb tumor suppressor gene is not limited to the rare childhood cancer from which it was isolated - it is also implicated in some of the most commonly occurring tumors of adults. As discussed in chapter 2, the Rb gene product also plays a central role in tumors induced by several DNA tumor viruses, particularly SV40, adenovirus, and human papillomaviruses. The transforming proteins of these viruses (T antigen, El A, and E7) all act, at least in part, by binding to the Rb protein and inactivating its function (Fig. 9.11). The inactivation of the Rb protein by binding to the oncogene proteins of these DNA tumor viruses has the same consequences as mutational inactivation of the Rb gene - uncontrolled cell proliferation and tumor development. Moreover, the interaction of these oncogene proteins with Rb provides a good illustration of the relation between oncogene and tumor suppressor gene functions, with an oncogene protein inducing cell transformation by inactivating the product of a tumor suppressor gene.
Rb Protein
- Controls transition from G1 S
- Braking system
- Mutations in Lung, Breast and Bladder CA
- Also caused by HPV
- Human cervcal cancer
- Oncogene for E7 binds and inactivates Rb
In the hypophosphorylated state, pRb is active and carries out its role as tumor suppressor by inhibiting cell cycle progression. Phosphorylation inactivates pRb. pRb is activated near the end of G1 phase when a phosphatase dephosphorylates one of its residues, allowing it to bind E2F.
