| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Human Cancer Biology |
Authors' Affiliations: Departments of 1 Genetics, 2 Clinical Epidemiology, and 3 Pathology, Faculty of Medicine, Memorial University of Newfoundland, St. John's, Newfoundland, Canada and 4 Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario, Canada
Requests for reprints: Michael Woods, Discipline of Genetics, Memorial University of Newfoundland, Health Sciences Centre, Room 4333, 300 Prince Philip Drive, St. John's, Newfoundland, Canada A1B 3V6. Phone: 709-777-7334; Fax: 709-777-7497; E-mail: mwoods{at}mun.ca.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Experimental Design: Seventy-eight colorectal cancer patients were accrued over a 2-year period from the Avalon Peninsula of Newfoundland. We also examined 31 hereditary nonpolyposis colorectal cancer–like families, which had been referred to the Provincial Medical Genetics Program. Tumors from probands were tested by immunohistochemistry for deficiencies in MLH1, MSH2, and MSH6 proteins and tested for DNA microsatellite instability. Mutation analyses of MLH1, MSH2, and MSH6 were undertaken by direct sequencing and an assay to detect deletions, amplifications, and rearrangements in MSH2 and MLH1.
Results: We identified eight population-based families that fulfill the Amsterdam I or II criteria, 4 (50%) of which seem to have hereditary cancer not attributable to the most commonly mutated mismatch repair genes. In addition, in 16 of 21 (76%) referred families fulfilling Amsterdam I or II criteria, no mutations were found in the three most commonly altered mismatch repair genes, and tumor analyses corroborated these findings.
Conclusions: It seems that strong and novel genetic causes of hereditary colorectal cancer are responsible for a high proportion of colorectal cancer in this population. Conditions are suitable for the identification of these genes by linkage studies of large Newfoundland cancer families.
Mutations in three mismatch repair (MMR) genes account for >95% of the known mutations causing HNPCC: MSH2 (14, 15), MLH1 (16, 17), and MSH6 (18). Several other MMR genes have been associated with HNPCC; however, their roles are not well established. Few kindred have been identified with germ line mutations in PMS1 (19, 20) or PMS2 (19, 21, 22), and there is no convincing evidence for involvement of MLH3 (23–25) or MSH3 in high-risk colorectal cancer families.
The Canadian province of Newfoundland and Labrador, with a current population of 
510,000, is considered a collection of genetic isolates (26, 27). Approximately 60% of the population live in communities of <2,500 inhabitants and 41% in communities of <1,000 people (28). Such a population structure and a willingness to participate in research studies have made Newfoundland and Labrador an invaluable resource for the study of genetic disorders. In fact, a large Newfoundland colorectal cancer kindred was used to identify the importance of MSH2 in HNPCC (14).
In the current study, we used a population-based study of colorectal cancer patients diagnosed in 1997 and 1998 to determine the proportion of hereditary colorectal cancer cases on the Avalon Peninsula of Newfoundland (Fig. 1) and to determine the genetic basis of disease in hereditary cases. We found a large number of families that seem to have a familial form of colorectal cancer, which cannot be explained by mutations in MMR genes. We then analyzed families referred to the Provincial Medical Genetics Program to determine if there are additional high-risk families in Newfoundland that have hereditary colorectal cancer not associated with MMR deficiencies. Our results suggest there are probable novel genetic causes of hereditary colorectal cancer in this distinct population and that conditions are suitable for the identification of these factors by linkage studies of large Newfoundland cancer families.
|
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
One hundred seventy-nine cases met the inclusion criteria. Contact was made with 158 (88%) eligible cases/proxies of which 106 (67%) agreed to participate in the study. The final number of subjects who provided family histories was 79 (44% of eligible cases).
Families referred to the genetics clinic. Over the last 20 years, >100 families have been referred to the Provincial Medical Genetics Program for assessment of possible hereditary colorectal cancer. Thirty-one families having the greatest family risk of HNPCC and with DNA samples available were examined in this study. All referred families had been screened previously and did not carry the most common Newfoundland founder HNPCC mutation (MSH2 c.942+3A>T; ref. 29). Informed consent was obtained from all subjects or an appropriate proxy. Ethics approval was granted by the Human Investigations Committee of the Faculty of Medicine, Memorial University of Newfoundland, the Health Care Corporation of St. John's and the Avalon Peninsula Health Board.
Microsatellite instability and immunohistochemistry. For the population-based probands, matched tumor and normal tissue from formalin fixed, paraffin-embedded blocks were compared for microsatellite instability (MSI) using at least five microsatellite markers. Markers used were from the National Cancer Institute panel (30). Individual PCR reactions for each of the markers were electrophoresed on 6% polyacrylamide gels and silver stained. Each case was designated as either MSI-high (MSI-H; 
30% markers unstable), MSI-low (<30% markers unstable), or microsatellite stable (MSS; no unstable markers; refs. 30, 31). When a MSI-low result was obtained, a second set of markers, including the mononucleotide BAT40 and four dinucleotide markers D17S787, D18S58, D20S100, and D7S519, were evaluated as above.
Colorectal tumors from the referred group underwent similar MSI analysis as above, with the exception that microsatellite markers were radioactively labeled and visualized by autoradiography (32). Only BAT25 and BAT26 were analyzed in the referred group, because in the population-based probands there was 98.6% (70 of 71) concordance between overall MSI-H and MSS status and the MSI status of these two mononucleotide microsatellites.
For the immunohistochemical analysis of MLH1, MSH2, and MSH6, formalin-fixed, paraffin-embedded tissues were sectioned at 4 µm, deparaffinized, and rehydrated with xylene and alcohol. The slides underwent either pressure cooker or microwave antigen retrieval [10 mmol/L citrate buffer (pH 6.0) for 3 minutes at 115°C in microMED T/T Mega; Hacker Instruments & Industries, Inc., Fairfield, NJ]. Nonspecific binding was blocked by 20% protein blocker with avidin (Signet Laboratories, Inc., Dedham, MA). The slides were washed with TBS. The sections were then incubated with mouse antibodies against MLH1 (1:40; G168-728, PharMingen, San Diego, CA), MSH2 (1:100; FE 11, Oncogene Research Products, Cambridge, MA), or MSH6 (1:100; 44, BD Transduction Laboratories, Mississauga, Ontario, Canada) for 1 hour. The antibodies were detected with the avidin-biotin complex method. 3,3'-Diaminobenzidine tetrahydrochloride was used as the chromogen, and hematoxylin was used as a counterstain.
Mutation detection. DNA from probands and, when available, additional family members was prepared from whole blood using a simple salting-out method (33). Alterations of MLH1, MSH2, and MSH6 were determined by direct sequencing of all 45 exons and intron/exon boundaries. DNA from all clinic-based patients and from those population-based probands who had tumors deficient in a MMR protein was subjected to direct sequencing. Automated sequencing was done on an ABI 377 DNA Sequencer (Applied Biosystems, Foster City, CA). Sequence information of the coding region was derived from RefSeq NM_000249.2 (MLH1), NM_000251.1 (MSH2), and NM_000179.1 (MSH6). Primer sequences and intronic nucleotide information were derived from genomic sequences from National Center for Biotechnology Information: AC011816.17 (MLH1), AC079775.6 (MSH2), AC006509.15 (MSH6). Primer sequences are available from the authors on request.
Exon deletions in MSH2 and MLH1 were detected by multiplex ligation–dependent probe amplification (MLPA) using DNA from all clinic-based probands and from those population-based probands whose tumors were deficient in MLH1 or MSH2 (34). MLPA, using the HNPCC probes (kit SALSA P003), was conducted and analyzed according to the protocol provided by Medical Research Council-Holland (Amsterdam, the Netherlands). All deletions identified in probands by MLPA were confirmed in other affected relatives by MLPA and/or reverse transcription-PCR from lymphocyte RNA.
All sequence variant nomenclature conforms to the recommendations found on the Human Genome Variation Society Web site (http://www.hgvs.org) updated February 27, 2005.
Detection of MLH1 promoter methylation. We used bisulfite modification coupled with methylation-specific PCR to determine methylation status of region C of the MLH1 promoter (35). Template DNA (1 µg) was treated as outlined in the Chemicon CpGenome Bisulfite Modification kit (Chemicon International, Temecula, CA). Primers and conditions for methylation-specific PCR were obtained from previously published work (36) and were examined by gel electrophoresis. Fully methylated DNA (Chemicon International) and normal blood DNA were used as controls.
Statistical analyses. Time to first cancer was analyzed using the Kaplan-Meier time to event analysis in first-degree and second-degree family members of probands. Low-risk families were used as the reference group for comparison between risk groups using Cox regression analysis. Probands were excluded from these analyses. In low-risk families, both parents and siblings of the parents were included, whereas in higher-risk families only the parent and siblings on the transmitting side were included.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
60 years. Eleven additional families fulfilled the ACMAC classification. The 32 probands/families that did not meet any of these criteria were classified as low risk.
Time to cancer in population-based families. The cumulative probability of developing colorectal cancer or any HNPCC-associated cancer (as defined by the revised Bethesda criteria) in family members at 50% risk was stratified by risk classification (Supplementary Tables S1 and S2). In ACI and ACII families, >50% of family members had developed a HNPCC-associated cancer by age 70 years. In families defined by ACMAC, the time to first cancer curve was parallel to that of members from ACI and ACII families, except that cancer developed 10 years later in life (Fig. 2A), with 26% developing the cancer by age 70 years. The observed relative risk of any HNPCC-related cancer in ACMAC families (9.9 times that of low-risk families) suggests a genetic predisposition for cancer in these families. Members of families defined by the revised Bethesda guidelines (excluding those defined by ACMAC) had a 3.9-fold increased risk of developing any HNPCC-related cancer compared with the low-risk families, with 14% developing the cancer by age 70 years. Forty-seven percent of ACI and ACII families and 18% of ACMAC families developed colorectal cancer by age 70 years. The risk of developing colorectal cancer was 31 times greater for ACI and ACII families and 9.9 times greater for ACMAC families compared with low-risk families (Fig. 2B). For those families defined by the revised Bethesda guidelines, the relative risk was 3.8 (compared with low-risk families), with 9.3% developing colorectal cancer by age 70 years.
|
|
|
Of the tumors from 22 probands who fulfilled revised Bethesda guidelines, 5 (22.7%) had tumors that were MSI-H (Table 2). The remaining 17 (77.3%) probands had tumors that were MSS and expressed all three MMR proteins.
Five of 32 (15.6%) tumors from probands of low-risk families were MSI-H and/or immunohistochemical deficient (Table 2). The remaining 27 (84.4%) probands from low-risk families had tumors that expressed all proteins and were MSS.
Mutation analyses of population-based probands. MLPA was implemented as the first screen for mutations in MLH1 and MSH2 in the population-based cases. A deletion of exon 8 [c.1277-?_1386+?del (p.Gly426_Gln462>GlyfsX5)] in MSH2 was identified in a single kindred using this technique (Table 2; Fig. 3). Subsequent DNA sequencing identified a single proband (study ID 122) carrying a previously identified founder mutation (29), affecting the splice donor site 3' of exon 5 in MSH2 (c.942+3A>T). One proband (00MG1598; Table 2), who had a tumor that was deficient in both MSH2 and MSH6, did not have an identifiable pathogenic alteration in MSH2 but did have a novel missense mutation in MSH6 [c.3415G>A (p.Gly1139Ser)]. In four probands with MSH6-deficient tumors, a causative alteration was not identified. A mutation was also not identified in the seven probands with tumors deficient in MLH1. Thus, definitive cancer causing mutations were found in only 2 of the 16 (12.5%) probands screened for sequence alterations.
Methylation analysis of MLH1 promoter. Seven tumors from population-based probands, which were MSI-H, immunohistochemical deficient, and with no mutation identified in MLH1, were tested to determine if their tumors showed methylation of MLH1. In all seven tumors, hypermethylation of the MLH1 promoter was evident (Fig. 3).
Description of the referred families. Thirty-one probands from families referred to the Provincial Medical Genetics Program were selected for HNPCC testing, 13 fulfilled the ACI, 8 fulfilled the ACII, 1 was categorized as ACMAC, and 9 satisfied the revised Bethesda criteria (see Tables 1 and 3).
|
Mutation analyses of referred cases. Using MLPA, two deletions were identified in four referred probands. A deletion of exon 8 of MSH2 was found in three probands, and a deletion of exons 4 to 16 inclusive [c.646-?_2802+?del (p.Ile216_Thr934)] was identified in a single proband. To our knowledge, this 13 exon deletion has not been documented previously. No exonic deletions or amplifications were found in MLH1. When tumors were available from probands who had an exonic deletion, they were MSI-H and deficient in MSH2 or both MSH2 and MSH6 (Table 3).
Regardless of MLPA, immunohistochemical, or MSI results, all probands underwent direct sequencing of exons and intron/exon boundaries to determine potential disease-causing mutations, to catalog known and novel polymorphisms, and to identify potential disease modifying alleles. In the two tumors, which were MSI-H but intact for all tested MMR proteins, analysis of genomic DNA was negative for mutations in the three MMR genes sequenced (Table 3). Another proband (no. 10737) carried a c.1787A>G (p.Asn596Ser) mutation, presumably causing the tumor to be MSH2 deficient and MSI-H.
Surprisingly, 11 of 17 (64.7%) probands from ACI and ACII families, with cancers that were MSS and intact for the MMR proteins, had no identifiable mutations in MLH1, MSH2, or MSH6. In addition, tumors from 9 of 10 (90%) probands fulfilling the ACMAC or revised Bethesda criteria were MSS and intact for all proteins tested. However, in one of these probands (no. 2275), a p.Ser144Ile substitution (c.431G>T) in MSH6 was detected, which segregates with the disease in this family.
Several other genomic variants were identified, some of which have not been reported previously and may represent alterations unique to the Newfoundland population (Supplementary Table S3).
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Additionally, there are many families from the population-based probands that have a strong history of HNPCC-related cancers not attributable to MMR deficiency and with MSS tumors. Tumors from 4 of the 7 (57.1%) available probands of ACI and ACII families and 17 of 22 (77.3%) tumors from probands fulfilling the revised Bethesda criteria do not have a MMR deficiency in the proteins analyzed. Many of these families had a large number of HNPCC-related cancers segregating in a pattern consistent with an autosomal dominant mode of inheritance and failed to meet Amsterdam criteria because of late onset of the cancers. The ACMAC incorporates the principles of ACII (10), but with an age cutoff of <60 years instead of 50 years, and includes all the cancers outlined in the revised Bethesda criteria (37). In this group of 11 families, the risk of colorectal cancer or any other HNPCC cancer was substantially higher than that in low-risk families or in those fulfilling revised Bethesda criteria. In addition, tumors from 4 (36.4%) probands of ACMAC families were deficient in a MMR protein and MSI-H. Tumors from the remaining 7 (63.6%) were MSS and intact for all proteins tested. Taken together, these findings indicate that the majority of families with a strong familial predisposition for HNPCC-associated cancers may have mutations in novel susceptibility genes. The likelihood that mutations in other MMR genes, such as PMS1, PMS2, MLH3, or MSH3, are causing susceptibility to cancer in these families is low. Very few families have been identified with MMR defects caused by heritable mutations in these genes (19, 22, 23, 40). However, the involvement of these loci cannot be dismissed as we have not screened for mutations in these genes.
Due to the high frequency of HNPCC-like families without MMR deficiencies identified in the population-based cohort, we investigated families referred to the Provincial Medical Genetics Program. In total, four families fulfilling ACI or ACII criteria segregated an exon 8 deletion in MSH2, causing a frameshift and a premature stop codon in exon 9 (p.Gly426_Gln462>GlyfsX5). A deletion of this same region was identified previously; however, the exact location of the breakpoints could not be determined as they fell within a 45-bp AluY consensus sequence, which is present in both intron 7 and intron 8 of MSH2 (41). We were able to map the breakpoints to this same 45-bp region in our probands (data not shown), suggesting that this alteration has a common etiology and could be a mutational hotspot, as there have been other published reports of the deletion of exon 8 in HNPCC families (32, 42). Genotyping our families with microsatellite markers is under way to determine, by haplotype analysis, if this is a founder mutation or represents mutations that have occurred independently due to a common mechanism. As well, a multiexonic deletion of MSH2, spanning exons 4 to 16 inclusive (p.Ile216_Thr934del), segregates in the very large Family 11. This kindred includes >80 mutation carriers that, like the families segregating the exon 8 deletion, would be good candidates for genotype-phenotype studies.
As in other studies that screen MMR genes for genomic alterations in HNPCC families, we have identified several missense mutations of unknown significance. In one proband with a MSI-H tumor deficient in MSH2 and MSH6 (proband no. 10737), a p.Asn596Ser substitution was found, which occurs in the MSH3/MSH6 interaction domain of MSH2. Although both asparagine and serine are uncharged polar amino acids and such a change is tolerated according to the sorting intolerant from tolerant algorithm (43), there has been documentation that this is a functionally important amino acid (44–47). In addition, it was absent from 192 control chromosomes. Thus, there is some evidence to suggest that this is the causative alteration in our family. Unfortunately, there is no other DNA available from this family to determine the segregation pattern of the variant.
In another proband (no. 45) with a tumor that is MSI-H and deficient in MSH2 and MSH6, there is no identifiable mutation in MSH2 but there is a p.Gly1139Ser alteration in MSH6. This variant is found in the P-loop of the ATP-binding domain required for hydrolysis of ATP (48). Sorting intolerant from tolerant analysis indicated that any substitution of this amino acid is predicted to affect protein function. In Saccharomyces cerevisiae, an alteration of the homologous residue (p.Gly987Asp) caused a decrease in ATP hydrolysis and the failure of MMR, but the MSH2-MSH6 complex showed greater binding affinity to the mispair in vitro (48). Therefore, the MSH6-MSH2 complex may be sequestered onto the heteroduplex DNA, without mending the mispair, and then targeted for degradation, thus causing the deficiency of both MSH2 and MSH6 proteins. Unfortunately, there were no other family members available for testing; however, this alteration was not observed in 192 control chromosomes.
Another proband (no. 2275), fulfilling the revised Bethesda criteria was shown to be heterozygous for p.Ser144Ile in MSH6. This substitution is a change from an uncharged polar to a nonpolar residue and occurs in a conserved amino acid. This missense variant has been identified previously in four families (49–51), none of which fulfilled ACI or ACII. One study also reported functional analysis using an S. cerevisiae model system (50). They determined that this change, which they did not identify in 199 normal control chromosomes, caused the loss of MSH6 function observed in their assay. We also did not observe this change in any of 192 control chromosomes. Additionally, our proband has two siblings with the same missense alteration who have had tubular adenomas at ages 38 and 53 years. In addition, two unaffected siblings without adenomas at age 53 years, undergoing regular colonoscopy screening, do not carry the alteration. The above evidence is consistent with this variant being the cause of colorectal cancer in our proband.
A final proband (no. 12536) was heterozygous for p.Val49Glu in MLH1 and had a colorectal cancer, which was MLH1 deficient and MSI-H. Additionally, his nephew, who was diagnosed with colorectal cancer at age 49 years, was MSI-H, MLH1 deficient, and had the same substitution. Two other cancer patients from the family also have this alteration. The Val49 residue falls within the ATPase domain of MLH1 and is a nonpolar, hydrophilic amino acid, whereas Glu49 is polar and hydrophobic. Recently, functional studies of the ATPase domain have been done on yeast-human hybrid constructs, indicating that this change impairs MMR (52). Additionally, no substitutions were observed at this amino acid in 192 of our control chromosomes.
In the four population-based probands with a MSH6-deficient tumor, no mutation was identified. The MSH6 alteration may include deletions too large to be detected by sequencing. It was not possible to perform MLPA for MSH6 due to the unavailability of the assay. In addition, if mutations were located deep within intronic regions as cryptic splice sites, or considerably upstream of the translation start site, their detection would not be possible.
The dearth of identifiable mutations in the referred families is noteworthy. In 16 of the 21 (76.2%) families fulfilling Amsterdam criteria, a mutation in MLH1, MSH2, and MSH6 could not be identified. Other investigators have also noted that a large fraction of high-risk families have no detectable mutations in the most commonly mutated MMR genes (53, 54). Such observations could be due to inadequate mutation detection methods. In addition, mutations in other colorectal cancer predisposition genes could be responsible for disease in our families. It has been shown that attenuated familial adenomatous polyposis and MYH-associated phenotypes may mimic HNPCC in that there are very few polyps present and tumors are MSS (55, 56). Therefore, defects in these genes cannot be excluded solely by phenotypic analysis.
However, when comparing probands with a MMR deficiency to those without, there was a significant difference between the colorectal tumor sites of each group (Table 4). Distal tumors were much more prevalent in non-MMR-deficient probands than in those who had a MMR defect. As well, unlike the MMR-deficient probands, there were no other cancer types detected in the non-MMR-deficient probands. As has been suggested previously, there seems to be important clinical differences between those families with MMR defects and those without (57, 58). We support the suggestion that there be another designation ("familial colorectal cancer type X") to describe those families without a MMR defect, suggesting that there is an unexplained cancer etiology (58). In addition, some recent studies have analyzed the molecular features of families not associated with MMR defects and have indicated that novel pathways toward cancer may be involved (59, 60).
|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: B.V. Bapat and P.S. Parfrey contributed equally to this work.
Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).
Received 4/ 4/05; revised 6/ 6/05; accepted 6/ 7/05.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Mrkonjic, S. Raptis, R. C. Green, N. Monga, D. Daftary, E. Dicks, H.B. Younghusband, P. S. Parfrey, S. S. Gallinger, J. R. McLaughlin, et al. MSH2 118T>C and MSH6 159C>T promoter polymorphisms and the risk of colorectal cancer Carcinogenesis, December 1, 2007; 28(12): 2575 - 2580. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. T. Campbell, L. Edwards, J. R. McLaughlin, J. Green, H. B. Younghusband, and M. O. Woods Cytochrome P450 17A1 and Catechol O-Methyltransferase Polymorphisms and Age at Lynch Syndrome Colon Cancer Onset in Newfoundland Clin. Cancer Res., July 1, 2007; 13(13): 3783 - 3788. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
