Joanne Stone

Immunohistochemistry for MSH2 and MHL1: a method for identifying mismatch repair deficient colorectal cancer

Colorectal cancers with DNA mismatch repair (MMR) gene mutations characteristically display a high rate of replication errors in simple repetitive sequences detectable as microsatellite instability (MSI). Most are the result of somatic MMR dysfunction; however, a subset are caused by germline mutations. The availability of commercial antibodies for MSH1 and MLH2 offers an alternative strategy to molecular methods for identifying MMR deficient cancers. To evaluate immunohistochemistry, MLH1 and MSH2 expression was studied using monoclonal antibodies in formalin fixed, paraffin wax embedded cancers. The immunohistochemical staining patterns of 23 cancers displaying MSI, including four cases with germline mutations, were compared with 23 microsatellite stable (MSS) cancers. All MSS cancers exhibited staining with both antibodies. Twenty two of the MSI cases showed absent MMR expression with either anti-MSH1 or anti-MLH2. The high sensitivity and predictive value of immunohistochemistry in detecting MMR deficiency offers a method of discriminating between MSI and MSS cancers caused by MSH1 and MLH2 dysfunction. The application and suitability of immunohistochemistry for the detection of MSI and as a strategy for prioritising the mutational analysis of MMR genes in routine clinical practice is discussed.

Low frequency of germline E-cadherin mutations in familial and nonfamilial gastric cancer

SummaryLittle is known about the relative contributions of genetic and environmental factors to the development of gastric cancer. Mutations in the cell adhesion molecule E-cadherin are recognized to be associated with the development of undifferentiated, diffuse and invasive gastric cancers. A recent study of two gastric cancer families has shown that germline mutations in the E-cadherin gene can be causative (Guilford P et al, Nature 1998; : 402–405). We have examined the E-cadherin gene for constitutive mutations in a systematic series of 106 gastric cancer patients, 10 with a family history of the disease and 96 sporadic cases. No pathogenic mutations were observed in any of the 106 patients. The results indicate that germline mutations in E-cadherin will not account for more than 3% of gastric cancers.

Prevalence of the APC E1317Q variant in colorectal cancer patients

The notion that some common variants of APC might confer an increased colorectal tumour risk is supported by studies of the I1307K polymorphism. Recently it has been proposed that the E1317Q variant is also associated with an increased risk. We have studied the prevalence of E1317Q in 364 colorectal cancer patients and in 290 controls. Two patients were shown to possess E1317Q. Neither had a family history of colorectal cancer or co-existent adenomatous polyps. Two controls also carried E1317Q. This finding suggests that E1317Q is unlikely to be associated with anything more than a moderate increase in risk of colorectal cancer.

Lobular carcinoma in situ of the breast is not caused by constitutional mutations in the E-cadherin gene

Lobular carcinoma in situ (LCIS) is an unusual histological pattern of non-invasive neoplastic disease of the breast occurring predominantly in women aged between 40 and 50 years. LCIS is frequently multicentric and bilateral, and there is evidence that it is associated with an elevated familial risk of breast cancer. Although women with LCIS suffer an increased risk of invasive breast disease, this risk is moderate suggesting that LCIS may result from mutation of a gene or genes conferring a high risk of LCIS, but a lower risk of invasive breast cancer. The high frequency of somatic mutations in E-cadherin in LCIS, coupled with recent reports that germline mutations in this gene can predispose to diffuse gastric cancer, raised the possibility that constitutional E-cadherin mutations may confer susceptibility to LCIS. In order to explore this possibility we have examined a series of 65 LCIS patients for germline E-cadherin mutations. Four polymorphisms were detected but no pathogenic mutations were identified. The results indicate that E-cadherin is unlikely to act as a susceptibility gene for LCIS.

Distinct Genetic and Epigenetic Changes in Medullary Breast Cancer

Genetic instability resulting in chromosome aneuploidy or mismatch repair defi ciency characterizes cancer. Medullary carcinoma (MC) of the breast is a specific form of breast cancer with unique clinical, epidemiologic, and prognostic features, sug gesting distinctive tumorigenic pathways. To investigate the nature of the genetic changes associated with MC we analyzed a series of 22 tumors. Chromosomal im balances were assessed by comparative genomic hybridization (CGH) and mismatch repair (MMR) deficiency tested for through assessment of microsatellite instability (MSI) and expression of MLH1 and MSH2 genes. MMR deficiency was detected in only a small proportion of cases. The chromosomal copy number changes showed some similarities to BRCAl-associated tumors. A high level of BRCA1 promoter hy permethylation was detected, suggesting a possible role of this gene in MC develop ment

Optimising methods for determining RER status in colorectal cancers

Approximately 13% of colorectal cancers display microsatellite instability (MSI), a form of replication error repair. Colorectal cancers developing in individuals with constitutional defects in the mismatch repair (MMR) genes hMLH1, hMSH2, hPMS1 and hPMS2 consistently show evidence of this phenomenon. Since MSI is indicative of MMR deficiency, testing colorectal cancers for MSI provides a method of refining the identification of carriers of germline MMR mutations. To assess which microsatellites represent the best reporters of replication error (RER) status we have examined 116 early onset colorectal cancers for MSI. MSI was assessed using eight dinucleotide- and two mononucleotide-repeat fluorescently labelled polymerase chain reaction (PCR) markers. The two mononucleotide repeat markers (BAT25 and BAT26) were highly sensitive and typing of either represents an efficient strategy for defining RER status of colorectal cancers and obviates the requirement of typing numerous microsatellite markers.

Mutations in Bcl10 are very rare in colorectal cancer

SummaryBcl10 is a recently identified gene reported to be involved commonly in human malignancy (Willis et al (1999) Cell 96: 1–20). To investigate whether it is frequently mutated in colorectal cancer we have analysed a series of 132 colorectal cancers and eight colorectal cancer cell lines for mutations in Bcl10. One feature of the Bcl10 gene is that it harbours two polyadenine tracts. These repeating elements in genes can be prone to a high rate of mutation if there is defective mismatch repair. To examine the possibility that Bcl10 may be preferentially mutated in mismatch repair-deficient cancers, 49 of the tumours and cell lines were known to be replication error (RER)-positive and, of these, ten were from individuals harbouring germline mutations in hMLH1 or hMSH2. No pathogenic mutations were detected in the tumours and only one mutation was found in the colorectal cancer cell lines. These results indicate that Bcl10 is unlikely to be involved in the pathways of colorectal carcinogenesis.

Analysis of Li-Fraumeni syndrome and Li-Fraumeni-like families for germline mutations in Bcl10.

The Li-Fraumeni syndrome (LFS) is a dominant disease whose hallmark is an increased risk of breast cancers, brain tumours, sarcomas, leukaemia and adrenal carcinoma. Some, but not all LFS and Li-Fraumeni-like (LFL) families are caused by TP53 mutations. Bcl10 is a recently identified tumour suppressor reported to be commonly mutated in a wide range of cancers. To investigate the possibility that Bcl10 is a susceptibility gene for LFS and LFL we have analysed 27 LFS/LFL families. No mutations were observed. This indicates that Bcl10 is unlikely to act as a susceptibility gene for LFS and LFL.

Contribution of germline MLH1 and MSH2 mutations to lobular carcinoma in situ of the breast.

Lobular carcinoma in situ (LCIS) is an unusual histological pattern of non-invasive neoplastic disease of the breast occurring predominantly in women aged between 40 and 50 years. LCIS is frequently multicentric and bilateral suggesting a genetic basis to the disease. The high frequency of microsatellite instability in lobular breast cancers, coupled with increased risk of breast cancer associated with germline mismatch repair gene mutations raises the possibility that mutations MSH2 or MLH1 might confer susceptibility to LCIS. To explore this possibility we have examined a series of 71 LCIS patients for germline MSH2 and MLH1 mutations. No mutations were detected in MSH2. Two sequence variants were identified in MLH1. The first was a CTT-->CAT substitution, codon 607 (exon 16) changing leucine to histidine. The other mutation detected in MLH1 was a TAC-->TAA substitution codon 750 (exon 19) creating a stop codon, predicted to generate a truncated protein. These findings suggest that mutations in MLH1 may underlie a subset of LCIS cases.

Allelic imbalance in colorectal cancer at the CRAC1 locus in early-onset colorectal cancer

A susceptibility gene to colorectal adenomas and carcinoma (CRAC1) on chromosome region 15q14 approximately q22 has been proposed on the basis of linkage in a single family. Allele-specific loss of heterozygosity (LOH) in tumors of affected family members suggests that the causative gene functions as a tumor-suppressor gene. The genes that are mutated in inherited cancer syndromes are often involved in the pathogenesis of sporadic cancer. To determine whether CRAC1 plays a role in colorectal carcinogenesis in general, we have studied 277 cases of early-onset colorectal cancers for allele loss at 15q14 approximately q22 using four microsatellite markers (D15S970, D15S117, D15S971, and D15S1028) that define the region of maximal linkage. The frequency of LOH detected was between 14% and 22%, but there was no significant association between LOH at each adjacent marker. Most cancers caused by loss of expression of a tumor suppressor involve large-scale deletion of one allele. On this basis, our findings suggest that CRAC1 is unlikely to be implicated in the development of colorectal cancer in general or, if involved, it is through small somatic mutations or other loss of function mechanisms rather than allele loss.