Introduction
Tumor suppressor genes were first identified by making cell hybrids between tumor and normal cells. On some occasions a chromosome from the normal cell reverted the transformed phenotype.
Several familial cancers have been shown to be associated with the loss of function of a tumor suppressor gene. The table below lists several of these syndromes.
A few of these tumor suppressor genes are described in more detail below. They include the retinoblastoma susceptibility gene (RB), Wilms' tumors (WT1), neurofibromatosis type-1 (NF1), familial adenomatosis polyposis coli (APC or FAP), and those identified through loss of heterozygosity such as in colorectal carcinomas (called DCC for deleted in colon carcinoma) and p53 which was originally thought to be a proto-oncogene. However, the wild-type p53 protein suppresses the activity of mutant alleles of p53 which are the oncogenic forms of p53.
back to the top
| Familial Cancer Syndrome | Tumor Suppressor Gene | Function | Chromosomal Location | Tumor Types Observed |
| Li-Fraumeni Syndrome OMIM data | P53 | cell cycle regulation, apoptosis | 17p13.1 | brain tumors, sarcomas, leukemia, breast cancer |
| Familial Retinoblastoma OMIM data | RB1 | cell cycle regulation | 13q14.1-q14.2 | retinoblastoma, osteogenic sarcoma |
| Wilms Tumor OMIM data | WT1 | transcriptional regulation | 11p13 | pediatric kidney cancer |
| Neurofibromatosis Type 1 OMIM data | NF1
protein = neurofibromin 1 | catalysis of RAS inactivation | 17q11.2 | neurofibromas, sarcomas, gliomas |
| Neurofibromatosis Type 2 OMIM data | NF2
protein=merlin or neurofibromin 2 | linkage of cell membrane to actin cytoskeleton | 22q12.2 | Schwann cell tumors, astrocytomas, meningiomas, ependymonas |
| Familial Adenomatous Polyposis OMIM data | APC | signaling through adhesion molecules to nucleus | 5q21-q22 | colon cancer |
| Tuberous sclerosis 1 OMIM data | TSC1
protein = hamartin | interacts with tuberin, exact function unknown | 9q34 | facial angiofibromas |
| Tuberous sclerosis 2 OMIM data | TSC2
protein = tuberin | GTPase activation of RAP1 and RAB5 | 16p13.3 | benign growths (hamartomas) in many tissues, astrocytomas, rhabdomyosarcomas |
Deleted in Pancreatic Carcinoma 4
Familial juvenile polyposis syndromeOMIM data | DPC4
also known as Smad4 | regulation of TGF-b/BMP signal transduction | 18q21.1 | pancreatic carcinoma, colon cancer |
| Deleted in Colorectal Carcinoma OMIM data | DCC | transmembrane receptor involved in axonal guidance via netrins | 18q21.3 | colorectal cancer |
| Familial Breast Cancer OMIM data | BRCA1 | repair of double strand breaks by interaction with Rad51 protein | 17q21 | breast and ovarian cancer |
| Familial Breast Cancer OMIM data | BRCA2 | similar to BRCA1 activity? | 13q12.3 | breast and ovarian cancer |
| Cowden syndrome OMIM data | PTEN | phosphoinositide 3-phosphatase
protein tyrosine phosphatase | 10q23.3 | gliomas, breast cancer, thyroid cancer, head & neck squamous carcinoma |
| Peutz-Jeghers Syndrome OMIM data | STK11
protein = serine-threonine kinase 11 | potential regulation of vascular endothelial growth factor (VEGF) pathway | 19p13.3 | hyperpigmentation, multiple hamartomatous polyps, colorectal, breast and ovarian cancers |
Hereditary Nonpolyposis Colon Cancer type 1
HNPCC1OMIM data | MSH2 | DNA mismatch repair | 2p22-p21 | colon cancer |
Hereditary Nonpolyposis Colon Cancer type 2
HNPCC2OMIM data | MLH1 | DNA mismatch repair | 3p21.3 | colon cancer |
| Familial diffuse-type gastric cancer OMIM data | CDH1
protein = E-cadherin | cell-cell adhesion protein | 16q22.1 | gastric cancer, lobular breast cancer |
| von Hippel-Lindau Syndrome OMIM data | VHL | regulation of transcription elongation | 3p26-p25 | renal cancers, hemangioblastomas, pheochromocytoma |
| Familial Melanoma OMIM data | CDKN2A
protein=cyclin-dependent kinase inhibitor 2A | cell-cycle regulation | 9p21 | melanoma, pancreatic cancer, others |
| Gorlin Syndrome OMIM data | PTCH
protein = patched | transmembrane receptor for sonic hedgehog (shh), involved in early development through repression of action of smoothened | 9q22.3 | basal cell skin carcinoma |
| Multiple Endocrine Neoplasia Type 1 OMIM data | MEN1 | unknown | 11q13 | parathyroid and pituitary adenomas, islet cell tumors, carcinoid |
back to the top
p53
Loss of heterozygosity at the short arm of chromosome 17 has been associated with tumors of the lung, colon and breast. This region of chromosome 17 includes the p53 gene.
The P53 gene was originally discovered because the protein product complexes
with the SV40 large T antigen. It was first thought that P53 was a dominant oncogene since cDNA clones isolated from tumor lines were able to cooperate with the
RAS oncogene in transformation assays. This proved to be a misleading since the cDNA clones used in all these studies were mutated forms of wild-type p53 and cDNAs from normal tissue were later shown to be incapable of
RAS co-transformation. The mutant p53 proteins were shown to be altered in stability and conformation as well as binding to hsp70.
Subsequent analysis of several murine leukemia cell lines showed that the P53 locus was lost by either insertions or deletions on both alleles. This suggested that wild-type p53 may be a tumor suppressor not a dominant proto-oncogene. Direct confirmation came when it was shown that wild-type p53 could suppress transformation in oncogene cooperation assays with mutant p53 and ras.
It has now been demonstrated that mutation at the P53 locus occurs in cancers of the colon, breast, liver and lung. Indeed, p53 involvement in neoplasia is more frequent than any other known tumor suppressor or dominant proto-oncogene!!!!
The protein encoded by P53 is a nuclear localized phosphoprotein. A domain near the N-terminus of the p53 protein is highly acidic like similar domains found in various transcription factors. When this domain is fused to the DNA-binding domain of the yeast GAL4 protein, the resulting chimera is able to activate transcription from genes containing GAL4 response elements. This suggests that p53 may be involved in transcriptional regulation. p53 has been shown to bind DNA, in vitro, that contains at least 2 copies of the motif 5'-PuPuC(A/T)(A/T)GPyPyPy-3'. This sequence motif suggests that p53 may bind to DNA as a tetramer. Binding as a tetrameric complex explains the fact that mutant p53 proteins act in a dominant manner. They are present in complexes with wild type p53 and alter the function of the normal tetramer.
Like pRB, p53 forms a complex with SV40 large T antigen, as well as the E1B transforming protein of adenovirus and E6 protein of human papilloma viruses. Complexing with these tumor antigens increases the stability of the p53 protein. This increased stability of p53 is characteristic of mutant forms found in tumor lines. The complexes of T antigens and p53 renders p53 incapable of binding to DNA and inducing transcription. A cellular protein, originally identified in a spontaneous transformed mouse cell line and termed MDM2, has been shown to bind to p53. Complexing of p53 and MDM2 results in loss of p53 mediated trans-activation of gene expression. Significantly, amplification of the MDM2 gene is observed in a significant fraction of most common human sarcomas.
Phosphorylation also appears to regulate the activity of p53. The level of p53 is low after mitosis but increases during G1. During S phase the protein becomes phosphorylated by cdc2 kinase and CKII. Sequences at the N-terminus of the p53 protein function as a transcription activator suggesting p53 may activate the transcription of genes involved in suppression of cell growth. Additionally, p53 protein has been shown to block the binding of DNA polymerase-a to SV40 large T, blocking replication of SV40 DNA. It is suggested that p53 may also regulate the initiation of DNA synthesis. If p53 is involved in both transcription and DNA replication then all mutants of p53 may affect these properties in different ways. This may account for why some mutants lose tumor suppressor activity while others behave as dominant oncogenes.
back to the top
More detailed information on p53
Retinoblastoma (RB)
In the familial form of this disease individuals inherit a mutant, loss of function allele from an affected parent. A subsequent later somatic mutational event inactivates the normal allele resulting in retinoblastoma development. This leads to an apparently dominant mode of inheritance. The requirement for an additional somatic mutational event at the unaffected allele means that penetration of the defect is not always complete.
In sporadic forms of tumors involving the RB locus 2 somatic mutational events must occur, the second of which must occur in the descendants of the cell receiving the first mutation. This combination of mutational events is extremely rare.
The locus of the RB gene, identified cytogenetically, is chromosome 13q14.1. A 4.7 kb RB transcript has been identified (by chromosomal walking and subsequent Northern blotting with genomic DNA probes) and subsequently cloned. The RB gene encompasses 27 exons that span 180 kb of chromosome 13. Two of the introns in this gene are extremely large, 35 kb and 70 kb. The RB RNA encodes a p110 kDa protein (pRB) of 928 amino acids. pRB is a nuclear localized phosphoprotein. pRB is not detectable in any retinoblastoma cells. However, surprisingly detectable levels of pRB can be found in most proliferating cells even though there is a restricted number of tissues affected by mutations in the RB gene (i.e. retina, bone and connective tissue).
Many different types of mutations occur to result in loss of RB function. The largest percentage (30%) of retinoblastomas contain large scale deletions. Splicing errors, point mutations and small deletions in the promoter region have also been observed in some retinoblastomas.
The germ line mutations at RB occur predominantly during spermatogenesis as opposed to oogenesis. However, the somatic mutations occur with equal frequency at the paternal or maternal locus. In contrast, somatic mutations at RB in sporadic osteosarcomas occur preferentially at the paternal locus. This may be the result of genomic imprinting.
pRB functions at some capacity in the regulation of cell cycle progression. Its ability to regulate the cell cycle correlates to the state of phosphorylation of pRB. Phosphorylation is maximal at the start of S phase and lowest after mitosis and entry into G1. Stimulation of quiescent cells with mitogen induces phosphorylation of pRB, while in contrast, differentiation induces hypophosphorylation of pRB. It is, therefore, the hypophosphorylated form of pRB that suppresses cell proliferation. One element in the growth suppressive pathway of pRB involves the c-MYC gene. Proliferation of keratinocytes by TGF-b is accompanied by suppression of
c-MYC expression. The inhibition of c-MYC expression can be abrogated by introducing vectors that express the SV40 and adenovirus large T antigens which bind pRB. Therefore, a link exists between TGF-b, pRB and
c-MYC expression in keratinocytes.
Transformation by the DNA tumor viruses, SV40, adeno, polyoma, human papilloma and BK is accomplished by binding of the transforming proteins of these viruses to pRB when pRB is in the hypophosphorylated (and thus the proliferation inhibitory) state.
back to the top
More detailed information on Retinoblastoma
Wilms Tumor (WT1)
Genetic evidence indicates that at least 3 distinct loci may be involved in the development of Wilms tumors, a kidney cancer found in children. Either one (unilateral) or both (bilateral) kidneys can be involved. Sporadic evolution of Wilms tumors is associated with chromosomal deletions, identified cytogenetically, at both 11p13 and 11p15. The 11p15 deletions may involve the IGF-2 or
c-Ha-RAS
loci. There are also familial forms of Wilms tumors that do not involve either locus. These studies suggest that lesions at 3 different loci could lead to Wilms tumors.
Only a single candidate locus has thus far been characterized. The potential Wilms tumor gene at 11p13 is found in a deleted region of about 345 kb. This region contains a single transcription unit identified as WT1 that spans 50-60 kb, contains 10 exons and specifies a 3-kb mRNA. The WT1 gene codes for a protein that is highly conserved evolutionarily suggesting an important role for this gene. Several functional domains have been identified in the WT1 protein. The first hint of the potential function for WT1 came from the identification of 4 zinc finger domains suggesting it to be a transcription factor. There is a potential leucine zipper motif in the center of the protein indicating that WT1 may associate other leucine zipper containing proteins. The NH2-terminal 180 amino acids are involved in self-association. Separate domains have also been characterized that are involved in autonomous repression and activation of transcription.
Expression of WT1 is very restrictive, unlike the RB gene which is expressed everywhere. Wilms tumors that are homozygous for 11p13 deletions do not contain WT1 mRNA. Most other Wilms tumors show high expression of WT1 mRNA but these are likely to be produced from mutated WT1 genes.
back to the top
More detailed information on Wilms Tumor
Neurofibromatosis Type 1 (NF1)
All cases of neurofibromatosis arise by inheritance of a mutant allele. Roughly 50% of all affected individuals carry new mutations which appear to arise paternally, possibly reflecting genomic imprinting.
Germ line
mutations at the NF1 locus result in multiple abnormal melanocytes (café-au-lait spots) and benign neurofibromas. Some patients also develop benign pheochromocytomas and CNS tumors. A small percentage of patients develop neurofibrosarcomas which are likely to be Schwann cell derived.
Assignment of the NF1 locus to chromosome 17q11.2 was done by linkage studies of affected pedigrees. The NF1 locus is extremely large as is the transcript encoded by the locus. The mRNA is 11-13 kb and contains a 7.5 kb coding region. The protein encoded is 2485 amino acids and shares striking homology to rasGAP. The NF1 protein has been given the name neurofibromin. Expression of NF1 is observed in all tissues thus far examined.
Development of benign neurofibromas versus malignant neurofibrosarcomas may be the difference between inactivation of one NF1 allele versus both alleles, respectively. However, changes other than at the NF1 locus are clearly indicated in the genesis of neurofibrosarcomas. A consistent loss of genetic material on the short arm of chromosome 17 is seen in neurofibrosarcomas but not neurofibromas. The losses at 17p affect the wild type p53 locus and may be associated with a mutant p53 allele on the other chromosome.
Characterization of the NF1 protein was carried out by generating antibodies against both fusion proteins and synthetic peptides. These antibodies specifically recognize a 220kDa protein, called neurofibromin, in both human and rat spinal cord. Neurofibromin is most abundant in the nervous system. Immunostaining of tissue sections indicates that neurons, oligodendrocytes, and nonmyelinating Schwann cells contained neurofibromin, whereas astrocytes and myelinating Schwann cells do not. In schwannoma cell lines from patients with neurofibromatosis, loss of neurofibromin is associated with impaired regulation of the GTP-bound form of the proto-oncogene RAS (GTP-RAS). Analysis of other neural crest-derived tumor cell lines showed that some melanoma and neuroblastoma cell lines established from tumors occurring in patients without neurofibromatosis also contained reduced or undetectable levels of neurofibromin, with concomitant genetic abnormalities of the NF1 locus. In contrast to the schwannoma cell lines, however, GTP-RAS was appropriately regulated in the melanoma and neuroblastoma lines that were deficient in neurofibromin. These results demonstrate that some neural crest tumors not associated with neurofibromatosis have acquired somatically inactivated NF1 genes and suggested a tumor-suppressor function for neurofibromin that is independent of RAS GTPase activation.
back to the top
More detailed information on Neurofibromatosis Type 1 - NF1
Familial Adenomatosis Polyposis (FAP)
Somatic mutations in the adenomatous polyposis coli (APC) gene appear to initiate colorectal cancer development in the general population, whereas it is germ line mutations that are responsible for familial adenomatous polyposis (FAP). The APC gene exhibits a dominant pattern of inheritance. Multiple colonic polyp development characterizes the disease. These polyps arise during the second and third decades of life and become malignant carcinomas and adenomas later in life. Genetic linkage analysis assigned the APC locus to 5q21. This region of the chromosome is also involved in nonfamilial forms of colon cancer. FAP adenomas appear as a result of loss-of function mutations to the APC gene. This is characteristic of tumor suppressors.
Identification of the APC gene was aided by the observation that 2 patients contained deletions at the locus spanning 100 kb of DNA. Three candidate genes in this region, DP1, SRP19 and DP2.5 were examined for mutations that could be involved in APC. The DP2.5 gene has sustained 4 distinct mutations specific to APC patients indicating this to be the APC gene. To date, more than 120 different germ line and somatic mutations have been identified in the APC gene. The vast majority of these mutations lead to COOH-terminal truncation of the APC protein.
The APC gene contains 15 exons spanning approximately 125 kb of DNA encoding an 8.5 kb coding region in the mRNA. Northern blotting detects an RNA of around 10 kb. An alternative form of exon 9 (9A) was also found that splices into the interior of exon 9 removing 101 amino acids from the full length APC transcript. The protein coding region of the APC gene is also extremely large encompassing 2844 amino acids. No similarities to known proteins was found except for several stretches of sequence related to intermediate filament proteins.
Using antibodies specific for the NH2-terminus of APC, it is possible to co-precipitate additional APC-associated proteins. One of these APC-associated proteins is b-catenin. The catenins are a family of proteins that interact with the cytoplasmic portion of the cadherins (cell-cell adhesion family of proteins), thus linking the cadherins to the actin cytoskeleton.
Catenins are equally important in the signaling cascade initiated by the Wnt family of proteins that are involved in embryonic patterning, development of the nervous system. The Wnt proteins are secreted factors that interact with cell-surface receptors. Wnt-receptor interaction induces the activity of the cytoplasmic phosphoprotein dishevelled. Activated dishevelled inhibits the serine/threonine kinase glycogen synthase kinase-3b (GSK-3b). When GSK-3b is inhibited, b-catenin becomes hypophosphorylated. The hypophosphorylated form of b-catenin migrate to the nucleus and interacts with transcription factors (in particular with T-cell factor/lymphoid enhancer-binding factor-1 (Tcf/LEF-1), thereby, inducing expression of various genes. The suspected role of APC in this pathway is to bind phosphorylated b-catenin. The APC-b-catenin complex stimulates the breakdown of b-catenin. Therefore, mutations which lead to a loss of APC, or to a loss of the portion of the APC protein that interacts with b-catenin, would lead to constitutive activation of Tcf/LEF-1 and unrestricted growth.
back to the top
More detailed information on APC
Deleted in Colon Carcinoma (DCC)
Loss of heterozygosity (LOH) on chromosome 18 is frequently observed in colorectal carcinomas (73%) and in advanced adenomas (47%), but only occasionally in earlier-stage adenomas (11 to 13%). The area of chromosome 18 which is observed to be lost resides between 18q21.3 and the telomere. A 370 kbp stretch of DNA from the region of 18q suspected to contain the tumor suppressor gene was cloned. Expressed exons were used as probes for screening cDNA libraries to obtain clones that encoded a gene which was given the name DCC ('deleted in colorectal carcinomas'). A YAC contig, containing the entire DCC coding region, has been characterized showing that the DCC gene spans approximately 1.4 Mbp and contains 29 exons
The expression of the DCC gene has been detected in most normal tissues, including colonic mucosa. Somatic mutations have been observed within the DCC gene in colorectal cancers. The types of mutations seen included a homozygous deletion of the 5' end, a point mutation within one of the introns, and 10 examples of DNA insertions within a 170 bp fragment immediately downstream of one of the exons.
Evaluation of sporadic colon cancers for allelic deletions defined an area of chromosome 18 that included two candidate tumor suppressor. One was DPC4 (see Table above) and the other was DCC. DPC4 is deleted in up to one-third of cases assayed and DCC, or a closely linked gene, was deleted in the remaining tumors. Tumor suppressor genes located on chromosome 17p and 18q are critically involved in the development of most gastric cancers. Involvement of DCC may be rather selective for gastrointestinal cancers. Loss of DCC gene expression is also an important factor in the development or progress of pancreatic adenocarcinoma.
The DCC protein is a transmembrane protein of the immunoglobulin superfamily and has structural features in common with certain types of cell-adhesion molecules, including neural-cell adhesion molecule (N-CAM). It is known that the establishment of neuronal connections requires the accurate guidance of developing axons to their targets. This guidance process involves both attractive and repulsive cues in the extracellular environment. The netrins and semaphorins are proteins that can function as diffusible attractants or repellents for developing neurons. However, the receptors and signal transduction mechanisms through which they produce their effects are poorly understood. Netrins are chemoattractants for commissural axons in the vertebral spinal cord. Recent work has shown that DCC is expressed on spinal commissural axons and possesses netrin-1-binding activity. This suggests that DCC is a receptor or a component of a receptor that mediates the effects of netrin-1 on commissural axons. Genetic evidence showing an interaction between DCC and netrin homologs in C. elegans (the UNC-40 protein) and Drosophila melanogaster (the frazzled protein) supports the role of DCC as a receptor in the axonal guidance pathway. Mice carrying a null allele of DCC harbor defects in axonal projections that are similar to those seen in netrin-1-deficient mice, further supporting the interaction between DCC and axon development. However, the DCC-deficient mice exhibited no effects on intestinal growth, differentiation or morphogenesis which fails to demonstrate a tumor-suppressor role for DCC.
DCC has been shown to induce apoptosis in the absence of ligand binding, but blocks apoptosis when engaged by netrin-1. Furthermore, DCC is a caspase substrate, and mutation of the site at which caspase-3 cleaves DCC suppresses the pro-apoptotic effect of DCC completely. DCC may function as a tumor-suppressor protein by inducing apoptosis during metastasis or tumor growth beyond the local blood supply, both of which are conditions that lack the DCC ligand. This would likely occur through functional caspase cascades leading to cleavage of DCC.
back to the top
Back to Topics<<<<
This article has been modified by Dr. M. Javed Abbas.
If you have any comments please do not hesitate to sign my Guest Book.
23:21 19/12/2002
|