Mechanisms of APC-driven tumorigenesis: lessons from mouse models.

Colorectal cancer still represents one of the most common causes of morbidity and mortality among Western populations. The adenomatous polyposis coli (APC) gene, originally identified as the gene responsible for familial adenomatous polyposis (FAP), an inherited predisposition to multiple colorectal tumors, is now considered as the true "gatekeeper" of colonic epithelial proliferation. It is mutated in the vast majority of sporadic colorectal tumors, and inactivation of both APC alleles occurs at early stages of tumor development in man and mouse. The study of FAP has also led to one of the most consistent genotype-phenotype correlations in hereditary cancer. However, great phenotypic variability is still observed not only among carriers of the identical APC mutation from unrelated families but also from within the same kindred. The generation of several mouse models carrying specific Apc mutations on the same inbred genetic background has confirmed the genotype-phenotype correlations initially established among FAP patients, as well as provided important insights into the mechanisms of colorectal tumor formation. Here we review the major features of the available animal models for FAP and attempt the formulation of a hypothetical model for APC-driven tumorigenesis based on the observed genetic and phenotypic variability in mouse and man.

The genetic background modifies the spontaneous and X-ray-induced tumor spectrum in the Apc1638N mouse model.

The effect of the genetic background on the tumor spectrum of Apc1638N, a mouse model for attenuated familial adenomatous polyposis (FAP), has been investigated in X-irradiated and untreated F1 hybrids between C57BL/6JIco-Apc1638N (B6) and A/JCrIBR (A/J), BALB/cByJIco (C) or C3H/HeOuJIco (C3). Similar to the ApcMin model, the Apc1638N intestinal tumor multiplicity seems to be modulated by Mom1. Moreover, several additional (X-ray-responsive) modifier loci appear also to affect the Apc1638N intestinal tumor number. The genetic background did not significantly influence the number of spontaneous desmoids and cutaneous cysts in Apc1638N. In general, X-irradiation increased the desmoid multiplicity in Apc1638N females but had no effect in males. The opposite was noted for the cyst multiplicity after X-rays. Surprisingly, X-irradiated CB6F1-Apc1638N females were highly susceptible to the development of ovarian tumors, which displayed clear loss of the wild-type Apc allele.

A cost-effectiveness analysis of colorectal screening of hereditary nonpolyposis colorectal carcinoma gene carriers.

BACKGROUND: It has been estimated that the prevalence of carriers of a mutated mismatch repair (MMR) gene among the general population in Western countries is between 5 and 50 per 10,000. These carriers have a risk of >85% of developing colorectal carcinoma (CRC) and therefore need careful follow-up. The objective of this study was to analyze the cost-effectiveness of CRC surveillance of carriers of a mutated MMR gene. METHODS: The authors constructed a model to estimate the potential health effects (life expectancy) and healthcare costs of two strategies: 1) surveillance, with colonoscopy every 2-3 years, and 2) no CRC surveillance. Estimates of the lifetime risk of developing CRC and the stage distribution of CRC for symptomatic patients were derived from the Dutch hereditary nonpolyposis colorectal carcinoma (HNPCC) registry. The CRC stage specific relative survival rates and the effectiveness of surveillance in preventing or detecting cancer early were based on Finnish studies. The costs of surveillance and treatment were derived from recent American studies. RESULTS: The results showed that 1) surveillance of gene carriers led to an increase in life expectancy of 7 years, and 2) the costs of surveillance under a wide range of assumptions are less than the costs of no CRC surveillance. CONCLUSIONS: CRC surveillance of HNPCC gene carriers appears to be effective and considerably less costly than no CRC surveillance and therefore deserves to be supported by governmental agencies and health insurance organizations.

Germline mutations in the 3' part of APC exon 15 do not result in truncated proteins and are associated with attenuated adenomatous polyposis coli.

Familial adenomatous polyposis (FAP) is an inherited predisposition to colorectal cancer characterized by the development of numerous adenomatous polyps predominantly in the colorectal region. Germline mutations in the adenomatous polyposis coli (APC) gene are responsible for most cases of FAP. Mutations at the 5' end of APC are known to be associated with a relatively mild form of the disease, called attenuated adenomatous polyposis coli (AAPC). We identified a frameshift mutation in the 3' part of exon 15, resulting in a stop codon at 1862, in a large Dutch kindred with AAPC. Western blot analysis of lymphoblastoid cell lines derived from affected family members from this kindred, as well as from a previously reported Swiss family carrying a frameshift mutation at codon 1987 and displaying a similar attenuated phenotype, showed only the wild-type APC protein. Our study indicates that chain-terminating mutations located in the 3' part of APC do not result in detectable truncated polypeptides and we hypothesize that this is likely to be the basis for the observed AAPC phenotype.

Multiple products in the protein truncation test due to alternative splicing in the adenomatous polyposis coli (APC) gene.

Reverse transcription-polymerase chain reaction (RT-PCR)-based analyses of the adenomatous polyposis coli (APC) gene encompassing exons 1-15 revealed a complex pattern of products that were due to alternative splicing of exons 9, 10A and 14. The multiplicity of polypeptide chains obtained from T7-promoter-directed in vitro translation of the RT-PCR product pool was confirmed immunochemically to correspond to the mRNA isoforms, but not to represent products of internal initiation of translation. This observation is of particular relevance for the diagnostic protein truncation test (PTT), since this assay will pick up mRNA variants derived from physiological splice events, e.g., skipping of exons 9, 10A and 14. In vitro-translated proteins of reduced molecular weight were therefore detectable in healthy individuals. We extended this observation to the PTT of cDNA encompassing APC exons 1-14 of familial adenomatous polyposis patients. Knowledge of the normal polypeptide pattern seen in the diagnostic in vitro translation assay allowed us not only to identify translational stop mutations, but even to detect a splice acceptor mutation of exon 14 as a result of quantitative changes of the isoform pattern. Western immuno blot analysis on protein extracts of Epstein-Barr virus-immortalized lymphocytes of the same patients revealed that mutations accessible to the RT-PCR PTT yield intracellularly undetectable APC proteins.

Molecular genetic tests as a guide to surgical management of familial adenomatous polyposis.

BACKGROUND: In familial adenomatous polyposis the only curative treatment is colectomy, and the choice of operation lies between restorative proctocolectomy (RPC) and colectomy with ileorectal anastomosis (IRA). The RPC procedure carries a higher morbidity but, unlike IRA, removes the risk of subsequent rectal cancer. Since the course of familial adenomatous polyposis is influenced by the site of mutation in the polyposis gene, DNA analysis might be helpful in treatment decisions. METHODS: We evaluated the incidence of rectal cancer in polyposis patients who had undergone IRA, and examined whether the requirement for subsequent rectal excision because of cancer or uncontrollable polyps was related to the site of mutation. FINDINGS: Between 1956 and mid-1995, 225 patients registered at the Netherlands Polyposis Registry had undergone IRA. In 87 of them, a pathogenetic mutation was detected. 72 patients had a mutation located before codon 1250 and 15 patients after this codon. The cumulative risk of rectal cancer 20 years after surgery was 12%, and at that time 42% had undergone rectal excision. The risk of secondary surgery was higher in patients with mutations in the region after codon 1250 than in patients with mutations before this codon (relative risk 2.7, p < 0.05). INTERPRETATION: On this evidence, IRA should be the primary treatment for polyposis in patients with mutations before codon 1250, and RPC in those with mutations after this codon.

The risk of brain tumours in hereditary non-polyposis colorectal cancer (HNPCC).

Hereditary non-polyposis colorectal cancer (HNPCC) is known to be associated with several extracolonic cancers, e.g., cancers of the endometrium, stomach, urinary tract, small bowel and ovary. An association between HNPCC and brain tumours has also been reported, although previous risk analysis did not reveal an excess of this type of tumour. To determine whether HNPCC predisposes patients to brain tumours, we used risk analysis to compare families with HNPCC to those in the general population. Of the 1,321 subjects from 50 HNPCC families (with 60,237 person-years of follow-up) in the Dutch HNPCC Registry which satisfy the Amsterdam Criteria, 312 had colorectal cancer. The registry revealed 14 brain tumours in the HNPCC-patients and their first-degree relatives: 5 astrocytomas, 3 oligodendrogliomas, 1 ependymoma and 5 tumours for which a pathological report was not available. The relative risk of brain tumour in patients with HNPCC and their first-degree relatives was 6 times greater than in the general population (95% confidence interval, 3.5 to 10.1). After exclusion of the cases based only on family history, the relative risk was 4.3 (95% confidence interval, 2.3 to 8.0). Although the relative risk of brain tumour was increased, the lifetime risk was low (3.35%). Because it is not certain whether an improvement of the overall prognosis can be achieved by early diagnosis and intervention, and in view of the low lifetime risk, we do not recommend screening for brain tumours in HNPCC families.

APC mutation in the alternatively spliced region of exon 9 associated with late onset familial adenomatous polyposis.

Germ-line mutations in the adenomatous polyposis coli (APC) gene are responsible for familial adenomatous polyposis (FAP). Genotype-phenotype correlation studies in patients with FAP have demonstrated associations of certain variants of the disease with mutations at specific sites within the APC gene. In a large FAP family, we identified a frameshift mutation located in the alternatively spliced region of exon 9. Phenotypic studies of affected family members showed that the clinical course of FAP was delayed, with gastrointestinal symptoms and death from colorectal carcinoma occurring on average 25 and 20 years later than usual, respectively. The numbers of colorectal adenomas differed markedly among affected individuals and the location of colorectal cancer lay frequently in the proximal colon. Our findings suggest that the exon 9 mutation identified in the pedigree is associated with late onset of FAP. The atypical phenotype may be explained by the site of the mutation in the APC gene. Analysis of the APC protein product indicated that the exon 9 mutation did not result in a detectable truncated APC protein. Given the location of the mutation within an alternatively spliced exon of APC, it is conceivable that normal APC proteins are produced from the mutant allele by alternative splicing.

Hereditary nonpolyposis colorectal cancer: results of long-term surveillance in 50 families.

A surveillance programme comprising either colonoscopy of sigmoidoscopy plus barium enema every 2-3 years was instituted in 50 hereditary nonpolyposis colorectal cancer (HNPCC) families. The families included 238 patients with colorectal cancer (CRC) (mean age at diagnosis: 43.7 years; range: 16-86 years). These patients had 597 first-degree relatives of whom 493 could be traced and 388 (79%) accepted the invitation for screening. The control group were relatives (index patients) with symptomatic CRC. The average follow-up duration was 5 years (1-20 years). Screening led to the detection of adenomas in 33 patients and CRC in 11 patients. Pathological examination revealed 1 Dukes' A, 7 Dukes' B and 3 Dukes' C cancers. In contrast, among the control group 47% had advanced CRC (Dukes' C or distant metastases). The 5-year survival of the screen-detected cases was 87% versus 63% in the control group. Of the 11 CRC cases in the screening group, 4 were detected within 1-4 years after a negative colonic examination. A large proportion of the polyps found in the screening and control groups showed a villous growth pattern and/or a high degree of dysplasia. We conclude that periodic examination of HNPCC families allows the detection of cancer at an earlier stage than in patients not under surveillance. Because of the possibly more aggressive nature of polyps associated with HNPCC, we recommend a screening interval of 1-2 years.

Age at diagnosis as an indicator of eligibility for BRCA1 DNA testing in familial breast cancer.

We searched for criteria that could indicate breast cancer families with a high prior probability of being caused by the breast/ovarian cancer susceptibility locus BRCA1 on chromosome 17. To this end, we performed a linkage study with 59 consecutively collected Dutch breast cancer families, including 16 with at least one case of ovarian cancer. We used an intake cut-off of at least three first-degree relatives with breast and/or ovarian cancer at any age. Significant evidence for linkage was found only among the 13 breast cancer families with a mean age at diagnosis of less than 45 years. An unexpectedly low proportion of the breast-ovarian cancer families were estimated to be linked to BRCA1, which could be due to a founder effect in the Dutch population. Given the expected logistical problems in clinical management now that BRCA1 has been identified, we propose an interim period in which only families with a strong positive family history for early onset breast and/or ovarian cancer will be offered BRCA1 mutation testing.

p53 mutations in colorectal cancers in the patients of Metropolitan New York.

An unselected series of 19 colorectal adenocarcinomas obtained from patients treated by surgery at Memorial Sloan-Kettering Cancer Center (MSKCC), New York, USA was investigated for mutations in exons 5, 6, 7 and 8 of the p53 gene. Ten of the tumors revealed at least one somatic mutation either in the exon 5 or 8. Two of them were found to carry two somatic mutations each. The DGGE pattern of one of these two tumors indicated that it contained two different clonal cell populations; a similar assessment was not possible in the other tumor. Sequence analysis of all the observed variants showed that eighty percent of the mutations were due to transitions and that half of them were at mutational hot spot codons, 175 and 273.

Human alpha B-crystallin (CRYA2) gene mapped to chromosome 11q12-q23.

The alpha B-crystallin gene (CRYA2) encodes the abundant lens protein alpha B-crystallin. A panel of human/rodent hybrid cell lines, derived from five different parental combinations, was characterized with respect to human chromosomal content and the presence of well-established human chromosome-specific markers. This panel was screened for the presence of CRYA2, using the third exon of the hamster alpha B-crystallin gene as a probe. The patterns of segregation of CRYA2 with individual human chromosomes show the highest degree of concordance between CRYA2 and chromosome 11. Using cell hybrids containing translocated and/or partially deleted human chromosomes, the CRYA2 gene was localized to 11q12-11q23.

Assignment of the human gene for histidine-rich glycoprotein to chromosome 3.

Histidine-rich glycoprotein (HRG) is a monomeric plasma glycoprotein involved in the modulation of coagulation and fibrinolysis. Using Southern analysis of human-rodent somatic cell hybrid DNA with a human HRG-specific cDNA probe, the HRG gene was assigned to chromosome 3. One hybrid that was known to contain only a segment of chromosome 3 also reacted positively with the HRG probe. Hybridization analysis with a set of chromosome 3-specific probes showed that the segment of chromosome 3 present in this hybrid is missing the region pter-p14, which indicates that HRG is not located in this region. No restriction fragment length polymorphisms were detected for HRG with 10 commonly used restriction enzymes.

Adenosine deaminase complexing protein (ADCP) expression and metastatic potential in prostatic adenocarcinomas.

The expression of the adenosine deaminase complexing protein (ADCP) was investigated by immunohistochemistry in the normal and hyperplastic human prostate, in 30 prostatic adenocarcinomas, and in seven human prostatic adenocarcinoma cell lines grown as xenografts in athymic nude mice. In the normal and hyperplastic prostate, ADCP was localized exclusively in the apical membrane and the apical cytoplasm of the glandular epithelial cells. In prostatic adenocarcinomas, four distinct ADCP expression patterns were observed: diffuse cytoplasmic, membranous, both cytoplasmic and membranous, and no ADCP expression. The expression patterns were compared with the presence of metastases. We found an inverse correlation between membranous ADCP immunoreactivity and metastatic propensity. Exclusively membranous ADCP immunoreactivity occurred only in non-metastatic tumours. In contrast, the metastatic tumours showed no or diffuse cytoplasmic ADCP immunoreactivity. This suggests that immunohistochemical detection of ADCP might predict the biological behaviour of prostatic cancer. However, the occurrence of membranous ADCP immunoreactivity in the xenograft of a cell line (PC-EW), derived from a prostatic carcinoma metastasis, indicates that not only the tendency to metastasize modulates ADCP expression.

The value of screening and central registration of families with familial adenomatous polyposis. A study of 82 families in The Netherlands.

In 1984 a national registry of families with familial adenomatous polyposis was set up in The Netherlands to promote screening in those families. Eight-two families had been registered by the end of 1988. Analysis of the pedigrees showed that 204 family members at risk had not yet been screened. The diagnosis of familial adenomatous polyposis was histologically confirmed in 230 patients. These patients were subdivided into two groups. Group A comprised patients with familial adenomatous polyposis referred because they were symptomatic, and Group B relatives of these patients who were found by screening to have familial adenomatous polyposis. The authors compared these groups with respect to the occurrence of colorectal carcinoma. Fifty-four patients were found to have a colorectal carcinoma at the time of diagnosis of familial adenomatous polyposis, i.e., 49 of the 104 patients in Group A (47 percent) and five of the 126 patients in Group B (4 percent). The average age at diagnosis of the 104 patients in Group A was 35 years (range, 13 to 66 years) and that of the 126 patients in Group B was 24 years (range, 8 to 59 years). By the age of 40 years, 90 percent of the patients in group B had been diagnosed. Late onset of familial adenomatous polyposis was found in four families. Endoscopy and/or radiography of the upper digestive tract were (was) performed in 44 of the 230 patients. Nineteen patients (43 percent) were found to have polyps in the stomach or duodenum, or both. In our series, only one patient died from cancer of the upper digestive tract (ampullary carcinoma). These results show conclusively that screening leads to the early detection of familial adenomatous polyposis. The value of a national registry is proved by the finding of many at-risk family members who had not previously been screened. Screening should start between the ages of 10 and 12 and should continue up to the age of 50. In the rare cases of families with an apparently late onset of familial adenomatous polyposis, screening should be continued up to age 60. More studies are needed to determine the natural history of polyps in the upper digestive tract.

Solid-phase adsorption of antigens for efficient production of antibodies reactive with native and fixed tissue antigens.

A study has been made of the efficacy of different immunization protocols using low antigen levels for the generation of monoclonal antibodies capable of detecting antigens (ADCP) in processed tissues. Protocols using unprocessed native antigen immobilized on nitrocellulose were more efficient than soluble antigen in generating serum antibodies reactive with both native antigen and processed tissues. The derived monoclonal antibodies reacted with native but not processed antigen. The use of antigen immobilized on polyvinylidene (PVDF) and subsequently processed as for histochemistry was successful in generating monoclonal antibodies reactive with processed antigen.

Facioscapulohumeral muscular dystrophy gene in Dutch families is not linked to markers for familial adenomatous polyposis on the long arm of chromosome 5.

Cosegregation of facioscapulohumeral muscular dystrophy (FSHD) and familial adenomatous polyposis (FAP) has been described in two small families. The gene for FAP is located on the long arm of chromosome 5. We studied two large Dutch families with FSHD and found no evidence for linkage with gene markers closely linked to FAP. These results strongly suggest that the FSHD gene segregating in the Dutch families is not localized close to the FAP locus on chromosome 5.

Isolation and characterization of cell hybrids containing human Xp-chromosome fragments.

We have subjected C12D, a Chinese hamster hybrid containing only the human X chromosome, to 6-thioguanine selection. The majority of the derivative clones retained rearranged Xp-fragments, which were characterized by using a combination of enzyme markers, DNA probes, and in situ hybridization. Two of these, TG2 and TG5sc9.1, contained only an Xpter----p21 fragment and should be an ideal resource for directed cloning from this region. A possible mechanism for the specific retention of Xp-fragments is discussed.

Localization of human adenosine deaminase (ADA) gene sequences to the q12----q13.11 region of chromosome 20 by in situ hybridization.

The gene for human adenosine deaminase (ADA), an enzyme constitutively expressed in all tissues investigated so far and deficient in some cases of severe combined immune deficiency, was previously assigned to chromosome 20 by syntenic analysis, using somatic cell hybrids and quantitative enzyme studies on patients with chromosome abnormalities. Attempts at regional localization of ADA through indirect approaches have so far resulted in uncertainties, as well as apparent inconsistencies. In situ hybridization of high-resolution somatic and pachytene chromosomes using a 3H-labeled cDNA probe of the ADA gene localized the gene to 20q12----q13.11. Rearrangements involving this region have been reported in various human hematological malignancies; in this regard, possible implications of the physical proximity of the ADA gene locus to that of SRC, an oncogene previously localized to the same region of chromosome 20, are briefly discussed.