Nature of disease
Lynch syndrome (LS) is the most common form of genetically defined, hereditary colorectal cancer (CRC), accounting for 1–3% of all such tumours. Historically, a variety of names have been used for the disease, originally identified by Aldred Scott Warthin in 1913 and then rediscovered by Henry T Lynch in 1966. Lynch coined the terms ‘site-specific colon cancer’ and ‘family cancer’ syndromes. During a workshop in Amsterdam in 1989, the participants agreed upon the name hereditary non-polyposis colorectal cancer (HNPCC), as at that time the syndrome was unknown to most doctors.1 The appropriateness of the name was discussed again at the international collaborative group on HNPCC meeting in Bethesda, MD, in 2004 where, as the syndrome is also associated with many other tumours, it was proposed that the name ‘Lynch syndrome’ should be reintroduced.1
Lynch syndrome is inherited as an autosomal dominant disorder, whereby if one parent has the disease, there is a 50% chance that each of his or her children will inherit it. It is characterised by an increased risk of CRC and cancers of the endometrium, ovary, stomach, small intestine, hepatobiliary tract, urinary tract, brain and skin among others, with the lifetime cancer risk highest for CRC ().
Lifetime cancer risk in LS
Overall, LS accounts for between 0.3% and 2.4% of CRCs, and its prevalence in the general population is of the order of 1 : 3100 (although this may be subject to underestimation due to the current lack of systematic testing).9,10 The risk of a second primary CRC in individuals with LS is high (estimated at 16% within 10 years) and the risk of a LS cancer in a first- or second-degree family member is approximately 45% for men and 35% for women by age 70 years.1
Lynch syndrome is caused by mutations in deoxyribonucleic acid (DNA) mismatch repair (MMR) genes, namely MutL homologue 1 (MLH1), MutS homologues 2 and 6 (MSH2 and MSH6) and postmeiotic segregation increased 2 (PMS2).4,11 Loss of DNA MMR activity in a cell, due to mutations in both alleles of one of the MMR genes, leads to an inability to repair base–base mismatches and small insertions and deletions, resulting in genetic mutations which may then progress to cancer.12 Mutations occur all over the genome, but especially in repetitive DNA sequences, such as microsatellites. These cause abnormal patterns of microsatellite repeats to be observed when DNA is amplified from a tumour with defective MMR compared with DNA amplified from surrounding normal tissue. This phenomenon is known as microsatellite instability (MSI).
Based on data from 12,624 observations worldwide, MLH1 accounts for 39%, MSH2 34%, MSH6 20% and PMS2 8% of entries in the International Society for Gastrointestinal Hereditary Tumours (InSiGHT) database (www.insight-group.org/mutations/). However, all such estimates are subject to bias, because these are generally mutations found in families referred to genetics clinics, subject to fulfilment of local referral guidelines.
Diagnosis/testing
Currently, the Amsterdam criteria (AC) II and Revised Bethesda criteria, as seen in , may be used to assist with diagnosis of LS. In 1989, the AC were proposed in order to provide uniform family material required for international collaborative research studies. In 1999, these criteria were revised to include extracolonic tumours.1 However, with the development of techniques to investigate tumours, such as MSI and MMR immunohistochemistry (IHC), in 1997 the Bethesda guidelines were developed to aid selection of tumours for testing and subsequently identifying individuals with LS. These guidelines were revised in 2004. It should be noted that all AC must be met whereas only one Bethesda criterion is necessary.
Criteria used to assist diagnosis of LS
The Bethesda criteria include MSI-high (MSI-H). This refers to MSI testing where the National Cancer Institute (NCI) has recommended a panel of five markers, known as Bethesda (or NCI) markers, which include two mononucleotides (BAT25 and BAT26) and three dinucleotide repeats (D2S123, D5S346 and D17S250). Tumours with no instability in any of the markers are considered to be microsatellite stable (MSS). When one reference marker is mutated, a tumour is considered to be MSI-low (MSI-L), and if two or more markers are altered, it is considered to be MSI-H.12 In some cases an additional panel of five markers is used; if 3 out of 10 show instability then it is classified as MSI-H, and if two or fewer, MSI-L.
Unfortunately, there are limitations to MSI testing due to MLH1 silencing commonly occurring in non-hereditary cancers. Thus, MSI is found in approximately 15% of sporadic CRC cases (i.e. CRC with no apparent hereditary component),1 and according to Umar and colleagues, as many as 50% of suspected cases of LS are not confirmed by a genetic defect (that is, mutation in one of the known MMR genes).12 Hence, the Bethesda criteria have been criticised as being insensitive and non-specific, because strictly applied they would result in approximately 25% of all CRC being tested. In turn, this has stimulated the development of additional tests for the diagnosis of LS, as presented in .
Overview of tests to assist with diagnosis of LS
Current evidence supports genetic testing for LS to include:13
evaluation of tumour tissue for MSI through molecular MSI testing and/or IHC of the four MMR proteins (MLH1, MSH2, MSH6 and PMS2)
molecular genetic testing of the tumour for
MLH1 gene methylation and/or somatic
BRAF V600E mutation to help identify those tumours more likely to be sporadic than hereditary, as the presence of a
BRAF V600E mutation makes LS very unlikely
1molecular genetic testing of the MMR genes to identify a constitutional (germline) mutation when findings are consistent with LS.
Prognosis
Colorectal tumours in LS appear to evolve through the adenoma–carcinoma sequence. However, this progression is accelerated compared with CRC in sporadic and other familial settings, i.e. 2–3 years as opposed to 8–10 years.12,14 Furthermore, adenomas in LS often occur in younger individuals and tend to be larger and more severely dysplastic than in sporadic cases.14 That said, recent studies have confirmed early suspicions that patients with CRC from LS families survive longer than sporadic CRC patients with same-stage tumours.15 The reasons for the favourable survival rate with CRC in this syndrome remain unclear, but are likely related to a reduced propensity to metastasise. Explanations include that immunological host defence mechanisms may be more active in tumours of the MSI Pathology Research International 3 phenotype, and that the relatively high mutational load that occurs in tumours with defective DNA repair systems is detrimental to their survival.14
Furthermore, there is definite evidence for a genotype–phenotype correlation in LS; for example, one study found that MSH6 mutation carriers had markedly lower cancer risks overall than MLH1 or MSH2 mutation carriers.2 Carriers of a MMR gene mutation have a very high risk of developing CRC (25–70%) and endometrial cancer (EC) (30–70%) and an increased risk of developing other tumours.5
Management of disease
Surveillance
As LS is a hereditary condition, identification of family members carrying a MMR gene defect is desirable, as colonoscopic surveillance, and possibly prophylactic and/or altered surgical management, may be offered to high-risk individuals.
Given that screening for a mutation is time-consuming and expensive – largely because four genes may have to be analysed and their mutational spectra are wide (Vasen 20071) – the British Society of Gastroenterology (BSG) and the Association of Coloproctology of Great Britain and Ireland (ACPGBI)9 recommend that individuals with a substantially elevated personal risk of gastrointestinal malignancy be offered surveillance on the basis of one or more of the following criteria:
a family history (FH) consistent with an autosomal dominant cancer syndrome
pathognomonic features of a characterised polyposis syndrome personally or in a close relative
the presence of a constitutional (‘germline’) pathogenic mutation in a CRC susceptibility gene
molecular features of a familial syndrome in a CRC arising in a first-degree relative (FDR).
Individuals fulfilling at least one of the above criteria should be referred to a NHS regional genetics centre for assessment, genetic counselling and mutation analysis of relevant genes, where appropriate.
Vasen and colleagues (2007)1 highlight a study in which 10-year surveillance of 22 LS families reduced the development of CRC by 60% and also decreased mortality.10,16 Appropriately targeted surveillance also means that those without a gene defect may be spared intensified surveillance, which is costly and carries not insignificant risks of morbidity and mortality.1
If LS has been identified, large bowel surveillance is recommended by the BSG and ACPGBI for probands and family members as follows:9
Total colonic surveillance (at least biennial) should commence at age 25 years. Surveillance colonoscopy every 18 months may be appropriate because of the occurrence of interval cancers in some series. Surveillance should continue to age 70–75 years or until co-morbidity makes it clinically inappropriate. If a causative mutation is identified in a relative and the consultand is a non-carrier, surveillance should cease and measures to counter general population risk should be applied.
Reproduced from Gut, Cairns SR, Scholefield JH, Steele RJ, Dunlop MG, Thomas HJ, Evans GD, et al., Volume 59, pp. 666–89, 2010 with permission from BMJ Publishing Group Ltd.
Families fulfilling Amsterdam criteria, but without evidence of DNA mismatch repair gene defects (following negative analysis of constitutional DNA and negative tumour analysis by microsatellite instability testing/immunohistochemistry), require less frequent colonoscopic surveillance.
Gastrointestinal surveillance should cease for people tested negative by an accredited genetics laboratory for a characterised pathogenic germ-line mutation shown to be present in the family, unless there was a significant, coincidental finding on prior colonoscopy.
The evidence for upper gastrointestinal surveillance in all of these disorders is weak, but limited evidence suggests it may be beneficial.
Reproduced from Gut, Cairns SR, Scholefield JH, Steele RJ, Dunlop MG, Thomas HJ, Evans GD, et al., Volume 59, pp. 666–89, 2010 with permission from BMJ Publishing Group Ltd.
Debate continues regarding the appropriate age for and frequency of surveillance, but the above criteria are in agreement with further published data.1 However, the situation becomes more complex when the proband does not have a detectable DNA alteration associated with LS, or when an alteration with an unclear significance is identified.4 Vasen and colleagues (2007)1 suggest that this is the case for approximately 30% of families meeting the AC I, for whom a less intensive surveillance protocol may be recommended (i.e. colonoscopy at 3–5 year intervals, starting 5–10 years before the first diagnosis of CRC or at > 45 years).
Surgical management
Several studies have shown that patients with LS have an increased risk of developing multiple (synchronous and metachronous) CRCs.1 The type of surgery received, i.e. total or subtotal colectomy, depends on the location of the tumour and the stage of the cancer. Studies have shown that adenomas in patients with LS are located mainly in the proximal colon (ascending and transverse);14 therefore, a subtotal colectomy is favoured, which involves removal of most of the colon, leaving a small amount to be reattached to the rectum. Clinicians may also discuss prophylactic colectomy as a reasonable option in mutation carriers for whom colonoscopy is painful or difficult, or for a patient with adenomas that cannot be removed easily; however, this remains controversial.
Chemotherapy
At least three chemotherapeutic agents have been proven to be effective in the treatment of CRC – 5-FU (also known as fluorouracil) with or without leucovorin (also known as folinic acid), oxaliplatin and irinotecan – although experimental and clinical studies suggest that MSI-H tumours are resistant to 5-FU-based chemotherapy. Therefore, according to Vasen and colleagues (2007),1 prospective clinical trials are needed before definitive recommendations can be given.
Epidemiological studies have demonstrated that non-steroidal anti-inflammatory drugs (e.g. aspirin) reduce the risk of CRC.14 A recent study showed that a daily dose of aspirin reduced the incidence of CRC in carriers of LS after 56 months’ follow-up.17 The mechanisms by which aspirin prevents the development of cancer are unknown, though some have suggested that aspirin may be proapoptotic in the early stages of CRC development. Importantly, the Colorectal Adenoma/Carcinoma Prevention Programme 2 (CAPP2) trial of aspirin prophylaxis in LS has demonstrated that aspirin treatment for up to 3 years reduces, a decade later, the overall incidence of LS-associated cancers, including CRC, by 63%.17 A further dosage determination trial (CAPP3) is therefore planned in LS patients worldwide (www.capp3.org) and highlights the importance of identifying individuals and families with LS.
Description of technologies under assessment
The major laboratory tests used in the evaluation of patients suspected of having LS include testing of tumour tissue using IHC, MSI testing and constitutional testing for MMR mutations (generally from peripheral blood mononuclear cells). Family members undergo predictive genetic testing for the pathogenic mutation identified in the proband (unless they have also developed a relevant cancer).4 Other tests which may be carried out on tumours include BRAF V600E and methylation of MLH1.
Immunohistochemistry
In families with an increased probability of a MMR gene mutation, IHC analysis for MMR proteins MSH2, MLH1 and MSH6 in tumour tissue may be used as the first step to confirm the presence of MMR deficiency. Pathogenic mutations in MMR proteins frequently lead to the absence of a detectable gene product, or expression of the protein in an abnormal location, for example in the cytoplasm rather than the cell nucleus. Therefore, when tumour tissue from patients suspected of having LS is stained for MMR proteins, a negative or less intense nuclear staining may be visible as compared with the surrounding normal colonic tissue used as a positive control.4,18
The advantage of IHC, as opposed to MSI, is that abnormal staining of a specific MMR protein is related to the underlying gene defect and can therefore direct further genetic mutation analysis.1 IHC is a well-established technique widely available in cell pathology laboratories; however, when used to analyse MMR proteins in the setting of LS diagnosis, it must be performed to an adequate standard. Hence, at a workshop in 2006 it was decided that MMR IHC should only be available within the NHS via a laboratory accredited to Clinical Pathology Accreditation standards, obliged to participate in the UK National External Quality Assessment Scheme Immunocytochemistry (NEQAS ICC) for MMR proteins.19 This workshop made a number of recommendations, including that MMR IHC should be performed for all four main MMR proteins, in part to address the issue of tissue fixation artefact. Care must also be taken in histopathological interpretation of MMR IHC that an adequate and representative tissue sample has been analysed.
Although MMR IHC can give useful and informative results, its sensitivity is limited by a number of factors, for example tissue fixation, the variety and different performance characteristics of primary antibodies and the fact that some pathogenic mutations may result in catalytically inactive but antigenically intact proteins.15,20–23 Hence, there is a place for MSI analysis in cases with a high prior probability of LS, but with apparently normal expression of the MMR proteins.1
A particular issue with IHC is that approximately 15% of sporadic colon cancers lose expression of MLH1 because of somatic hypermethylation of the gene’s promoter. Therefore, whereas abnormal expression of MSH2, MSH6 or PMS2 is in itself reasonably good evidence that a tumour was due to LS, loss of MLH1 in itself is not. Other evidence must be used in interpretation in these circumstances, and thus testing for BRAF V600E and/or MLH1 promoter methylation may also be performed. The presence of the BRAF V600E mutation indicates a sporadic rather than LS-associated CRC, but the absence of BRAF V600E does not distinguish between sporadic tumours and those caused by LS. Similarly, MLH1 promoter methylation is highly correlated with a sporadic origin for a tumour, but is not absolutely conclusive, because individuals and families are described with constitutional MLH1 promoter methylation defects.24
Microsatellite instability testing
Microsatellite instability refers to the variety of patterns of microsatellite repeats observed when DNA is amplified from a tumour with defective MMR compared with DNA amplified from surrounding normal colonic tissue. Repetitive mono- or dinucleotide DNA sequences (microsatellites) are particularly vulnerable to defective MMR.4 MSI is prevalent in tumours from patients with MMR mutations, and in patients meeting either AC.4 Therefore, microsatellite analysis is commonly used as the first diagnostic screening test for LS.18
Microsatellite instability testing involves amplification of a standardised panel of DNA markers (Bethesda/National Institutes of Health markers), although laboratories may use 10 or more markers and, more recently, a commercially available kit based on five mononucleotide markers has become popular as mononucleotide microsatellites may be the most sensitive markers for use in detecting MSI.4 The process involves microdissection of tumour tissue, followed by extraction of DNA which is then amplified and run on a DNA fragment length analyser. Using such microsatellite markers, additional peaks in tumour tissue DNA in comparison with normal tissue DNA indicate MSI.25 Instability in 30% or more of the markers is considered MSI-H, less than 30% MSI-L and no shifts or additional peaks MSS. However, if instability is observed at any mononucleotide markers, MSI may be diagnosed. For this reason, MSI testing is moving to a smaller panel of mononucleotide markers, making the process more efficient and cheaper.
As for any molecular pathological analysis, tissue to be selected for MSI analysis must be first assessed by a histopathologist, prior to some degree of microdissection, which aids in maximising sensitivity. There is debate regarding the relative costs of MMR IHC and MSI testing, but NHS service laboratory costings indicate there is little to choose between the two. MSI may be more reproducible and can be performed with smaller amounts of tissue.26 As there is not yet a UK NEQAS scheme for MSI, the reproducibility of MSI compared with IHC is not established.
BRAF V600E and methylation testing
The presence of MSI in the tumour by itself is not sufficient to diagnose LS because 10–15% of sporadic CRCs exhibit MSI.25 MSI in non-LS tumours is usually caused by hypermethylation of the MLH1 gene. This acquired epigenetic inactivation of MLH1 is typically associated with mutations in the BRAF gene (specifically the V600E mutation), which has been described in ≈ 35% of sporadic MSI-H CRCs.25 Therefore, identification of hypermethylation of MLH1 and/or BRAF V600E is an indication that a patient does not have the LS germline mutation.
Ideally, tests would be performed together as the presence of the BRAF V600E mutation theoretically reduces the chance of LS as the cause of that tumour; however, because any test has a finite false negative (FN) rate, it is still a possibility. Additionally, if MLH1 promoter methylation is present but the BRAF V600E mutation is not, this would highlight the small possibility that the patient may have LS due to a constitutional MLH1 methylation defect. It is also possible that he or she could have an inherited MLH1 genetic mutation and could have acquired MLH1 promoter methylation as the ‘second hit’ in the tumour. In these cases, loss of heterozygosity of chromosome 3p (where MLH1 is located) is observed.25
Constitutional genetic testing
Multiple methods have been used for constitutional genetic testing (tests for mutations that affect all cells in the body and have been there since conception) in LS, in order to find inherited or, if de novo, potentially inheritable MMR gene mutations. The method(s) used should ideally be able to detect any possible mutation associated with LS, for example nonsense, missense and frameshift mutations, genomic deletions, duplications and rearrangements, as explained in and .4
Mutation types associated with LS
Measuring the accuracy of diagnostic tests for Lynch syndrome
One aspect of the evaluation of new tests is measuring their accuracy by calculating their sensitivity and specificity. This requires specification of the best available method of identifying the target condition of interest, known as the reference standard. Most mutations causing LS are point mutations or small insertions or deletions, suitably detected by DNA sequencing. However, some LS-associated mutations are deletions/duplications of exons in MLH1 and MSH2. These are more difficult to detect and, currently, the most appropriate technology available is multiplex ligation-dependent probe amplification (MLPA), which is a multiplex polymerase chain reaction (PCR) method able to simultaneously detect copy number changes across multiple DNA sequences within one sample. Therefore, the ideal reference standard is considered to be sequencing plus MLPA.
