Diagnosing Colorectal Cancer and its Limitations
Colon cancer is the second leading cause of cancer death. Most of the cases are sporadic but several inherited familial syndromes account for around 5% of all colorectal cancers (CRC). The most common of these syndromes are familial adenomatous polyposis (FAP) and hereditary non-polyposis colon cancer (HNPCC) which is also known as Lynch syndrome. These make up 3% of the familial syndromes. Other syndromes include Peutz-Jeghers syndrome and Juvenile polyposis syndrome. This essay will focus on the most common hereditary CRC's.
The genes that are involved in FAP and HNPCC have been identified. Therefore genetic testing can be offered to screen for carrier status in these syndromes. Identification of individuals who have a predisposition to CRC is important to be able to offer them adequate screening to detect tumours at an early stage.
FAP is dominantly inherited and is highly penetrative showing a vertical transmission through a family pedigree. Its classical phenotype involves the growth of hundreds of adenomatous polyps along the colonic mucosa. If the polyps are not removed there is near a 100% chance of colon cancer. The average age of polyp development and colon malignancy is 16 and 39 years respectively. ‘FAP is found in ~1 per 7-10,000 births in the United States population and accounts for less than 1% of all CRC' (Davidson, 2007). FAP is a monogenetic disease and is caused by the mutation or deletion of the adenomatous polyposis coli (APC) gene which is found on chromosome 5. ‘95% of APC mutations that lead to FAP is either nonsense (28%) or truncating frameshift (67%)' (Burt and Neklason, 2005). The remaining 5% is caused by large deletions or rearrangements. ‘The APC gene is a tumour-suppressor gene and the APC protein is part of the Wnt-signalling pathway, involved in cell growth control' (Burt and Neklason, 2005). Mutations in the APC gene therefore cause activation of the Wnt-signalling pathway and uncontrolled cell growth.
There is another variant of FAP known as Attenuated form of FAP (AFAP), it has a later age on onset (>40), less adenomatous polyps (<50) and a lower risk of CRC. Some of these patients will have a mutation in the extreme of the 3' or 5' end of the APC gene compared to those who have extreme polyposis where mutations tend to be in the mid-portion of exon 15. It is important to be able to distinguish between the types of FAP to know where to screen for mutations and how to treat the condition.
It is also important to consider a patient's nationality when they present to clinic. This can determine their inheritance risk and help to locate the mutation. For example Ashkanazi Jews have a high prevalence of the I1307K mutation resulting in a lifetime risk of CRC between 10-20%.
Genetic testing is important in families who are at risk of FAP due the dominant and high penetrance phenotype. As there is a classical phenotype, FAP is easier to diagnose in clinic. In FAP nearly all mutations consist of truncations of the APC protein. This occurs by point mutations, causing either a frameshift by an insertion or deletion, or a nonsense codon. Genetic Testing for FAP is done by indicating the presence of a disease causing mutation by a Protein Truncation Test (PTT). Then the location of the mutation on the APC gene is found by Conformation-Specific Gel Electrophoresis (CSGE), Single-Strand Conformation Polymorphism (SSCP) or Denaturing Gradient Gel Electrophoresis (DGGE). Once the location of the mutation is localised the APC gene is sequenced to identify the disease-causing mutation. For those large deletions and rearrangements, Southern Blotting, Array CGH and MLPA can be used for identification. If all of these methods fail to identify the mutation, linkage testing to the APC gene can be done. As it has become cheaper to sequence the APC gene recently many clinics skip the process of locating the area of mutation and sequence the whole gene.
The APC gene is large and the mutations spread along it. This can make it difficult to locate the mutation. As many families have unique mutations it may be difficult to decide if the mutation found is actually pathogenic. In fact in patients presenting with phenotypical classical FAP, known pathogenic mutations are only found in 85% of them. The rest, although they may have changes in the APC gene it is unknown what these changes mean. As this information is uninformative it is not possible to give patients a risk of getting CRC or to screen their families to be able to exclude those who are not at risk.
All of the molecular tests used for FAP have their advantages and disadvantages. PTT fails to detect truncations that occur at the very end or beginning of a gene and large mutations. Additionally it cannot detect missense mutations. However, if it does find a mutation it is always disease causing. CSGE does detect more than 90% of mutations present. SSCP detects between 60%-95% of mutations and DGGE can detect up to 90% of sequence changes. Array CGH will miss small deletions and MLPA cannot detect balanced translocations and is sensitive to impurities. Therefore some mutations are being missed.
Linkage analysis can be between 90%-95% effective in families that have multiple members affected by the disease. The results of linkage give asymptomatic family members the risk they have of carrying the mutation. However, if these risks are not below 5% or above 95% they are not very useful in clinic. Furthermore not all families will have multiple affected members to be able to carry out linkage. ‘Additionally reduced protein expression may give rise to disease but causative mutations can be very difficult to find because they may be in regulatory areas' (Burt and Neklason, 2005).
If a patient presented to clinic with FAP phenotype but when tested no mutation could be found on the APC gene it is worth testing the MYH gene for mutations. The phenotype of MYH-associated polyposis (MAP) is similar but less severe than FAP and it is inherited recessively. It is important to distinguish between these different types so that the mutations can be identified so other members of the family can be screened. It is also important to treat the patient accurately.
‘25% of cases of FAP arise as spontaneous APC mutations' (Davidson, 2007). Only children of these patients would be at risk of being a carrier of the mutation. However as these patients will have no family history an inexperienced clinician may not recognise the condition as being FAP.
Individuals with HNPCC have an increased risk of developing CRC. It is the most common form of inherited CRC accounting for between ‘3-5% of all cases' (Davidson, 2007). It is an autosomal dominant condition and people who inherit the condition have early onset of colon cancer (<40years). Tumours develop mainly in proximal colon and a person affected ‘often has family history of colon cancer or other associated cancers such as endometrial, ovarian, brain, small intestinal, pancreatic and urinary tract' (Davidson, 2007). However there is no distinctive form of phenotype for HNPCC. It is important though, due to the higher lifetime risk of developing these cancers, to diagnose families affected, so that preventative screening can be offered.
The first step to diagnose HNPCC is when a patient presents in clinic. A criteria known as the Amsterdam criteria was devised over 15 years ago which must be met for an individual to be clinically diagnosed. This method has been criticised as being too rigid. There are now adopted versions of this criterion, such as the Modified Amsterdam and Modified Bethesda. These other criteria are more inclusive but less specific for HNPPC.
When a family is identified as potentially having HNPCC, they are eligible for diagnostic screening. HNPCC causes a higher risk of CRC due to a germline mutation of a mismatch repair (MMR) gene. The cell is then unable to process DNA repair. ‘Mutation carriers exhibit a characteristic phenotype termed microsatellite instability, characterised by expansion or contraction of short repeat sequences of DNA at multiple loci' (Syngal et al, 1999). ‘Pathogenic mutations have been found at four mismatch repair genes (MSH2, MHL1, PMS2 and MSH6), but so far most HNPCC cases are caused by mutations in either MLH1 or MSH2' (Muller et al., 2004).
High microsatellite instability is a signal that the MMR gene is deficient. ‘A tumour is considered MSI high or unstable if more than 40% of the loci show instability' (Burt and Neklason, 2005). If an MMR-high reading is found, a further diagnostic test, Immunohistochemical staining can be used to identify which gene is most likely to be mutated. This looks for the MMR proteins MSH2, MLH1 and MSH6 in the tumour tissue.
In 50%-70% of cases, mutations in mismatch repair genes can be found by DNA sequencing and the larger deletions and rearrangements which tend to be common in HNPCC can be found by Southern Blotting. Southern Blotting will find the mutation in a further 10%-20% of people, where sequencing could not.
In clinic a diagnosis is dependent on a patient's family history. If the family history fits the Amsterdam or any of the modified criteria then they can be identified as potentially having HNPPC. However this method of diagnosing has potential flaws. A detailed family history may not be given by the patient as they may not be in contact with other family members or they may leave out information that they do not think is vital, such as endometrial cancer. The sensitivity of the Amsterdam criteria is between 54%-91% and the specificity is between 62%-84%. This means that a substantial number of HNPCC families could be excluded from testing and screening.
The sensitivity of MSI tests are 62%. Therefore extra caution should be used when interpreting results, especially negative results. In 15% of sporadic cases of CRC, MSI can be detected. ‘This occurs due to methylation of the 5'CpG island in the promoter region' (Muller et al., 2004). The result of this test may lead you to believe that the proband has HNPPC and therefore them and their families would be at higher risk of developing CRC. However this is not caused by a heritable mutation in the germline. In addition this ‘phenomenon exposes a corresponding limitation in the use of IHC because MLH1 protein expression is lost in tumours as well' (Lynch et al., 2007)
Additionally, reliable results for MSI can only be obtained if enough cells are correctly amplified to look at the microsatellite loci.
These tests do not pick up all cases of HNPCC as ‘approximately 10% of IHC tests will be falsely negative, i.e. protein stain is present even though the related gene is mutationally inactivated' (Burt and Neklason, 2005). This could lead to the prevention of early detection. Diagnostic treatment can also give false positive results; this can lead to people receiving screening that do not need it and this may lead to unnecessary psychological stress. MSI and IHC tests complement each other and therefore both should always be taken into account.
When the gene is identified that is likely to be mutated, sequence analysis or Southern Blotting can be performed to identify the mutation. However this result is not always informative for families as it can be difficult to clearly define a pathogenic mutation. This means that it is not possible to screen other family members for a known mutation. Also in up to 10% of people a mutation may not be found.
Finally when choosing an index case to look for the mutation, the youngest affected person should be chose. The older a patient is the more likely their cancer is sporadic. However in a family the youngest affected person may not agree to be tested. This leaves it harder to achieve the results wanted.
The most difficult part of testing for CRC is knowing who to test. In FAP there is a clear phenotype but there are variants such as AFAP and MAP. These also benefit from testing. HNCPP has no clear phenotype and is much harder to gauge who would benefit from testing. There are criteria set in place to help this process but as mentioned some people are over looked.
For both conditions no single test is sufficient to identify the mutation. Some mutations may not even be identified and therefore it is important that a consultant not only assess the patient on their test results but also on their clinical diagnosis. Test results are not required for disease management but are helpful for identifying other family members who are at risk.
Burt, R. & Neklason, D., W. (2005). Genetic Testing for Inherited Colon Cancer. Gastroenterology 128 : 1696-1716
Davidson, N., O. (2007). Genetic testing in colorectal cancer:who, when, how and why. Keio Journal of Medicine 56(1) : 14-20
Lynch, P., M. (2007). New Issues in Genetic Counseling of Hereditary Colon Cancer. Clinical Cancer Research 13 : 6857s-6861s
Muller, A., Giuffre, G., Edmonston, T., B., Heinmoller, E., Brodegger, T., Tuccari, G., Mangold, E., Buettner, R., & Ruschoff, J. (2004). Challenges and Pitfalls in HNPCC Screening by Microsatellite Analysis and Immunohistochemistry. Journal of Molecular Diagnostics 6(4) : 308-315
Syngal, S., Fox, E., A., Li, C., Dovidio, M., Eng, C., Kolondner, R., D., & Garber, J., E. (1999). Interpretation of Genetic Test Results for Hereditary Nonpolyposis Colorectal Cancer: Implications for Clinical Predisposition Testing. Journel of the American Medical Association 282(3) : 247-253
Boultwood, J. & Fidler, C. (2002). Molecular Analysis of Cancer, Humana Press, New Jersey.
Chen, S., Watson, P. & Parmigiani, G. (2005). Accuracy of MSI Testing in Predicting Germline mutations of MSH2 and MLH1: a case study in Bayesian meta-analysis of diagnostic tests without a god standard. Biostatistics 6(3) : 450-464
Kerr, D., J., Young, A., M. & Hobbs, F., D., R. (2002). ABC of Colorectal Cancer, BMJ, London.
Taylor, I., Garcia-Aguilar, J. & Goldberg, S., M. (2002). Colorectal Cancer second edition, Health Press, London
Tomlinson, I., Ilyas, M. & Novelli, M. (1997). Molecular Genetics of Colon Cancer. Cancer and Metastasis Reviews 16 : 67-69
Walther, A., Johnstone, E., Swanton, C., Midgley, R., Tomlison, I. and Kerr, D. (2009). Genetic Prognostic and Predictive Markers in Colorectal Cancer. Nature Reviews 9 : 489-499