Info

* see OMIM database for details on these syndromes

* see OMIM database for details on these syndromes

Figure 13.2 The APC gene and its encoded protein A: Organization of the gene as shown in the ensembl data base. There is a second, alternative promoter with a non-coding additional first exon ┬╗15 kb upstream of exon 1 shown here. B: Sequence features and functional domains of the APC protein. C: Localization of germ-line (nonsense) mutations in familial colorectal cancers. Point mutations in sporadic cases are most prevalent in the region affect in Gardner syndrome.

Figure 13.2 The APC gene and its encoded protein A: Organization of the gene as shown in the ensembl data base. There is a second, alternative promoter with a non-coding additional first exon ┬╗15 kb upstream of exon 1 shown here. B: Sequence features and functional domains of the APC protein. C: Localization of germ-line (nonsense) mutations in familial colorectal cancers. Point mutations in sporadic cases are most prevalent in the region affect in Gardner syndrome.

binding several different proteins and of oligomerization. The oligomerization domain is located near the N-terminus. Several repeats are recognizable along the length of the protein, of which the catenin-repeats and the 20 aa-repeats are involved in assembling a protein complex containing P-Catenin, glycogen synthase 3p (GSK3P), and the scaffold protein Axin or its relative Conductin (or Axin2). This protein complex is involved in regulating WNT signaling (see below). A basic region in APC located further towards the C-terminus interacts with microtubules. The actual C-terminal region binds EB1 and DLG, two proteins interacting with the mitotic spindle. There is some evidence that one function of APC lies in chromosome segregation. Loss of this function may favor the development of aneuploidy in colon cancer.

Patients with FAP inherit a mutation in one APC allele. Most are nonsense mutations leading to a truncated or instable protein and almost all are located in the first half of the gene. As predicted by the 'Knudson' model (^5.1), adenomas and carcinomas either contain somatic mutations in the second allele as well or have lost the intact allele by 5q deletion or by recombination, as recognizable by LOH in the 5q21 region.

Not unexpectedly, the gene mutated in the Gardner and Turcot syndromes as well as in the attenuated variety of FAP is also APC. The various syndromes are in general distinguished by the location of the mutation, which appears to determine the spectrum of organs in which hyperplasia and tumors arise. For instance, truncating mutations between codons 463 and 1387 (cf. Figure 13.2) are associated with the congenital hypertrophy of the retinal pigment epithelium characteristic of standard FAP, while mutations between codons 1403 and 1578 are often found in families with Gardner syndrome. Interestingly, truncating mutations within the first 150 aa cause attenuated FAP, i.e. a milder phenotype. Nonsense mutations closer to the translational start would be expected to usually cause a complete loss of a protein, while mutations in later exons would be thought to more frequently yield truncated, but stable proteins. It is therefore conceivable that the truncated protein produced by mutations further downstream in the APC gene may somehow aggravate the disease, perhaps by interfering with the function of the normal APC protein encoded by the intact allele. Although these mechanistical relationships are still incompletely understood, cataloguing of such genotype-phenotype relationships is important to optimize the treatment and counseling of the affected patients and families.

Since deletions and LOH of chromosome 5q are among the most frequent alterations in colorectal cancer overall, following the identification of APC as the gene mutated in FAP, sporadic carcinomas were screened extensively for mutations in the gene. Today, it is assumed that both alleles of the gene are inactivated by point mutation and deletion/recombination, or occasionally promoter hypermethylation, in 70-80% of colon and rectal cancers, irrespectively of whether they are familial or sporadic. Moreover, the frequency of APC mutations is almost the same in early and late stage tumors. Thus, APC is a prototypic tumor suppressor. Since its inactivation appears to be almost mandatory for the development of colorectal tumors and most probably takes place at an early stage, the designation of 'gatekeeper' is appropriate (^5.4).

The crucial role of APC is underlined by the analysis of the 20% or so colon cancers, which retain a fully functional APC protein at normal expression levels. Almost all of these contain mutations in other components of the APC/p-Catenin/Axin/GSK3p complex. Most often, mutations in the CTNNB1 gene encoding P-Catenin are found. More precisely, then, it is a disturbance of WNT signaling (cf. 6.10) that is so crucial for development of colorectal cancers.

There are >23 different WNT proteins in man. They are usually produced and secreted by mesenchymal tissues and act on neighboring epithelial cells in a paracrine manner. Most are agonists, but some may be antagonists. In the canonical pathway (Figure 13.3), WNT factors bind to one of several cell surface receptors named Frizzled (FZD). These receptors belong to the large 'serpentine' class of receptors characterized by seven transmembrane helices and interaction with trimeric G proteins. Activation of a FZD receptor activates the DSH (dishevelled) protein which in turn inhibits GSK3p. In a normal cell, this protein kinase is assembled together with Axin and APC in a cytosolic protein complex. This complex binds P-Catenin. In the absence of a WNT signal P-Catenin is phosphorylated by GSK3P at several sites near its N-terminus. This phosphorylation allows P-Catenin to be recognized by a ubiquitin ligase complex in which PTRCP performing the actual recognition.

Inhibition of GSK3P as a consequence of a WNT signal leads to accumulation of P-Catenin. The protein migrates into the nucleus, where it binds and activates TCF transcription factors, specifically TCF4 in colon cells. There are four known members of the TCF family, TCF1, LEF1, TCF3, and TCF4, which in the inactive

Figure 13.3 Alterations of the WNT pathway in colon cancers The WNT pathway is most often activated in colon cancers by mutations in APC (prohibition sign) that leads to disaggregation of the complex phosphorylating P-Catenin. In other cases, P-Catenin mutations (exploding star) prevent its inactivation. Constitutive activation of the pathway may often be exacerbated by loss of SFRPs that normally compete with FRPs for co-binding of WNT factors. See also Figure 6.12.

Figure 13.3 Alterations of the WNT pathway in colon cancers The WNT pathway is most often activated in colon cancers by mutations in APC (prohibition sign) that leads to disaggregation of the complex phosphorylating P-Catenin. In other cases, P-Catenin mutations (exploding star) prevent its inactivation. Constitutive activation of the pathway may often be exacerbated by loss of SFRPs that normally compete with FRPs for co-binding of WNT factors. See also Figure 6.12.

state are complexed with transcriptional repressor proteins, usually from the Groucho family. This repression is alleviated by P-Catenin and transcription of TCF target genes resumes. Among these target genes are several that act directly on cell proliferation and survival, such as CCDN1 and MYC.

Before becoming implicated in WNT signaling, P-Catenin had been known for a long time as a cytoskeletal protein binding to E-Cadherin which mediates homotypic cell interactions (^9.2). Therefore, the concentration of active P-Catenin is also modulated by E-Cadherin. This may allow to integrate information from WNT signals and cell adhesion (^9.2). Several further factors modulate WNT activity, e.g. SFRPs (secreted frizzled-related proteins), which interfere with WNT binding to receptors, and LRPs which support WNT binding to FZD receptors.

This intricate network is fundamentally disturbed in colon cancer (Figure 13.3). Obliteration of APC function, the most frequent alteration, causes a permanent WNT signal in the nucleus, since the regulatory protein complex cannot be assembled and P-Catenin does not become phosphorylated. As a result P-catenin accumulates and causes a constitutive activation of TCF target genes. Mutations in CTNNB1 alter the amino acids in the recognition sequence of GSK3P prohibiting phosphorylation of P-Catenin, with essentially the same consequence. The rarer deletions of Axin likewise disturb efficient phosphorylation of P-Catenin by removing the platform on which the proteins interact. Inactivating mutations have also been reported for PTRCP impeding the ubiquitin ligase that initiates the proteolytic breakdown of P-Catenin. These central alterations in the WNT pathway may be compounded by alterations in modulating factors. During tumor progression many carcinomas lose E-Cadherin expression which may exacerbate the accumulation of P-Catenin. Likewise, WNT signaling may be further enhanced by down-regulation of SFRPs.

Many other cancer types besides colorectal cancer have meanwhile been investigated for mutations in APC and its interacting genes. Two important conclusions can be drawn.

(1) Mutations in components of the WNT signaling pathway are found in many other carcinomas, with varying frequencies, although in none, it appears, with the same regularity. High to moderate prevalences are observed in hepatoma (^16.2), medulloblastoma, and breast cancer, lower frequencies in prostate carcinoma, renal carcinoma and glioblastoma. Of note, some of these cancers are also more frequent in the context of the Gardners and Turcots syndromes. Clearly, however, some cancer types lack mutations in the pathway; although it is difficult to exclude the occurrence of mutations with current techniques. Moreover, one promoter of the APC gene tends to hypermethylated in many carcinomas; it is not clear, however, whether this an indication of differential promotor use or overall diminished transcription.

(2) While APC is the major target in colorectal cancer, in other cancers mutations in CTNNB1 predominate. It is not at all clear, why this is so. One possibility is that only colorectal carcinogenesis requires obliteration of other APC functions as well, such as the one presumed in mitosis. Another possibility is that the difference is related to the mechanisms involved in carcinogenesis in different

Apoptosi┬╗

Proliferation

WNT WNT

Figure 13.4 Organization of proliferation and differentiation in colon crypts See text for detailed explanation.

organs. Distinct carcinogens involved may preferentially target certain positions in certain genes.

The requirement for constitutive activation of the WNT pathway in colon cancer may be closely related to the organization of this tissue (Figure 13.4). The colon epithelium is a constantly renewing tissue. The cells of the colon epithelium are derived from a small number - maybe five or so - of tissue stem cells located near the bottom of the crypt in a stem cell niche (^8.6). When these cells divide asymetrically, one daughter cell retains the stem cell character while the other is committed to differentiation. It moves up the crypt, gradually differentiating and concomitantly losing its ability to proliferate. Cells that have arrived at the surface are sloughed off or die by apoptosis.

It appears that the WNT pathway is central to the regulation of this renewal strategy. WNT factors are produced by mesenchymal cells near the stem cell niche allowing reproduction and maintaining immortality of the stem cell population. As cells committed to differentiation move away from the growth factor source, differentiation and apoptosis programs are turned on. Cell-associated signaling proteins called Ephrins and their receptors are involved in establishing the differentiated state and are down-regulated by WNT signaling. Thus, colon epithelial cells not exposed to WNT signals may enter a default state of differentiation or apoptosis. Constitutive WNT signaling caused by APC loss of function or P-Catenin over-activity would prohibit differentiation and apoptosis and establish a stem-cell like state independent of position in the tissue.

A similar effect is thought to be exerted by activation of the Hedgehog pathway in basal cell carcinoma of the skin (^-12.3). It is tempting to speculate that the range of cancers in which mutations of the WNT pathway are observed may be related to the proliferation strategy of the respective tissues. There is still too little data on this question, though, for a judgement on this hypothesis.

0 0

Post a comment