Available Assays MSI

MSI is the phenotypic reflection of an underlying MMR defect in a tumor. For this analysis, both tumor and normal tissue or cells are required. Typically, paraffin-embedded tissue sections (10 |im thick) are microdissected to obtain relatively pure tumor cells. Areas containing >70% tumor cells are typically used, although regions containing as little as 50% can provide adequate results. Normal cells are derived from adjacent normal mucosa, sections from archival normal surgical specimens for the same patient, if available, or peripheral blood. DNA is extracted separately from the tumor and normal cells. A number of extraction methods are now available that provide good-quality DNA from paraffin-embedded tissue. Isolated DNA is analyzed by the polymerase chain reaction (PCR) using radioactive or fluorescent-labeled primers to amplify and visualize multiple segments of DNA containing microsatellite loci. The PCR products are subjected to size-based separation on polyacrylamide gels or, more recently, through capillary electrophoresis.

MSI can be observed by comparing the electrophoretic patterns of amplified DNA from both tumor and normal tissue and is scored as the presence of novel fragments in tumor DNA compared to normal DNA (Figure 19-1a). For dinucleotide repeats, these novel fragments can be seen as higher- or lower-molecular-weight fragments (expansion or contraction of repeat). For mononucleotide repeats, however, lower-molecular-weight fragments are typically detected. Allelic imbalance, or LOH, is not scored as MSI since this alteration is the result of an entirely different mechanism (Figure 19-1, panel b). Overall, we have found that the most important aspect of this test is to have a well-optimized PCR assay. Because of PCR artifact, results will be far more difficult to interpret if each of the assays is not sufficiently optimized.

Historically, a major obstacle to understanding the tumor phenotype associated with defective MMR was a lack of consensus on the definition of MSI. The number of microsatellite loci and the proportion that demonstrate instability in order to be classified as MSI positive were highly variable between studies. The likelihood that any particular microsatellite will be susceptible to instability in a tumor probably relates to the inherent instability at the locus. Thus, differences in marker susceptibility to instability may lead to differences in the number of tumors scored as MSI positive.

In 1998, guidelines proposed for MSI testing at a National Cancer Institute (NCI) workshop were pub lished.19 These guidelines provide a reference panel of five validated microsatellite loci to ensure uniformity of diagnosis between laboratories. Laboratories could use the panel directly or use this panel as a basis of comparison to validate their own set of markers. These loci consist of two mononucleotide (BAT 25 and BAT 26) and three dinu-cleotide (D2S123, D5S346, and D175250) repeat markers. Mononucleotide markers are relatively stable and tend to be monomorphic. Mononucleotide markers were included in the panel because reports indicate that cells with mutant hMSH6 were mainly associated with instability at mononucleotide markers, while cells with mutations in hMLHl, hMSH2, or hPMS2 lead to pronounced instability at both mono- and dinucleotide repeats.48 Thus, analysis of only dinucleotide markers could lead to false-negative MSI results for tumors caused by mutations in hMSH6.

Several studies have reported that the fraction of markers demonstrating instability within tumors have a bimodal distribution with a breakpoint at approximately 30% to 40%.49 Therefore, tumors analyzed for MSI can be divided into three discrete groups: MSI-H (high-frequency MSI), MSI-L (low-frequency MSI), and MSS (microsatellite stable). For the NCI panel of five markers, MSI-H is defined as two or more unstable markers, MSI-L as one unstable marker, and MSS as no unstable markers. Although this group of five markers (or a comparable group) often is adequate to identify tumors with the MSI-H phenotype, it is not particularly useful for distinguishing between MSS and MSI-L. A larger number of markers is necessary to make this distinction. When only one marker is unstable or when technical difficulties limit the number of informative markers, an additional five markers may be necessary to help differentiate between MSI-H and MSI-L/MSS. In these cases, the absolute number may not be as important as defining the overall frequency of unstable microsatellites. When more than five markers are used, instability at >30% of markers is considered MSI-H, while MSI-L is defined as instability in <30% of the markers tested. More recent data suggest that the use of a greater number of mononucleotide markers (two to four per set), by themselves, may have a greater sensitivity and specificity for the detection of the MSI-H phenotype.50 Additionally, because these mononu-cleotide markers tend to be monomorphic, investigators have suggested that testing can be performed with tumor only (i.e., without the need to compare to normal tissue from that patient).51 However, caution should be used with this approach since other alleles can be detected in some patients, especially in certain racial groups. Without the use of normal tissue from that patient, such results could be scored as false positives. An example of a patient that is heterozygous for BAT 26 is shown in Figure 19-1c. In addition, the quasimonomorphic marker BAT26, which is located immediately downstream of exon 5 in the MSH2 gene, is widely regarded the most sensitive and specific marker of MSI. Pastrello and colleagues recently demonstrated that in individuals with large germline deletions of MSH2 which span exon 5, BAT26 is more likely to show no

85 9 0 95 100 105 110

05-Patient1-Tumor.fsa mor.fsa 6 Yel aaJLjjvaA!

4 Yell

155 1 60 ,165.170 1 75_180 185 190 195 200 205 210


11-Patient1-Tumor.fsa 9 Green mycL

Î2000 1500 1000 500

06-Patient1-Normal.fsa 4 Yellow

12-Patient1-Normal.fsa 11 Green

1000 500

1000 500

01-Patient1-Tumor.fsa 6 Blue

BAT 26

100 1 05 110 115 1 20 125 130 135 140 145 150 155 1 6!

01-Patient1-Tumor.fsa 6 Yellow

6 Yellow D5S346 : 9

06-Patient1-Normal.fsa 8 Blue

02-Patient1-Normal.fsa 8 Yellow

3000 2000 :1000

2000 1500 1000 500

03-Patient1-Tumor.fsa 5 Blue

03-Patient1-Tumor.fsa 5 Yellow

BAT 40



BAT 34c4 2000 1000

04-Patient1-Normal.fsa 7 Blue

04-Patient1-Normal.fsa 7 Yellow rmal.fsa 7 Blue

3000 2000 1000

140' 145 "150 155 160 1 65 170_175 180 185 190 I0?105.110"5 120 125 130 140 145 150 155 160

07-Patient1-Tumor.fsa 10 Blue 01-Patient1-Tumor.fsa 6 Yellow ient1-Tumor.fsa


8-Patient1-Normal.fsa 12 Blue

02-Patient1-Normal.fsa 8 Yellow

11-Patient1-Tumor.fsa 9 Blue

03-Patient1-Tumor.fsa 5 Yellow

2000 1500 1000

12-Patient1-Normal.fsa 11 Blue

D10S197 213000000000

1 Blue

BAT 25 3000 2000 1000

04-Patient1-Normal.fsa 7 Yellow

2000 1500 1000 500

3000 2000 1000

95 100 105 110 115 120 1 25 130 135

05-Patient5-Tumor.fsa 10 Green a 10 Green


06-Patient5-Normal.fsa 12 Green

900 600 300

90 95 100 105 110 115 120 1 25 130 135 140 1 45

07-Patient6-Tumor.fsa 10 Green


3000 2000 1000

08-Patient6-Normal.fsa 12 Green

2000 1500 1000 500

115 1 20 125 130 135 1 40 145 150 155 1 60 1 65 1 70 175

9-Patient7-Tumor.fsa 6 Yellow


3000 2000 1000

10-Patient7-Normal.fsa 8 Yellow


2000 1500 1000

3000 2000 1000

2000 1500 1000 500

3000 2000 1000

2000 1500 1000

BAT 26

85 90 95 100 1 05

10 115 1 20 125 130

01-Patient2-Tumor.fsa Tumor

1 Blue À

Heterozygous pat'ent

4000 3000 2000 1000

02-Patient2-Normal.fsa Normal

3 Blue


2000 1000


5 Blue


4000 2000


7 Blue


2000 1000

Figure 19-1. MSI analysis of normal and tumor DNA from a CRC patient using four mononucleotide repeat markers (BAT 25, BAT 26, BAT 40, BAT 34c4) and six dinucleotide repeat markers (D17S250, ACTC, D10S197, D18S55, D5S346, mycL) with the use of an

ABI3100 genetic analyzer. (a) Example of MSI for each of the ten markers. (b) Example of microsatellite changes that can be misinterpreted as MSI: allelic imbalance or LOH.(c) Example of a patient that is heterozygous for the marker BAT 26.

140 145

MSI when tested on tumor tissue, due to the subsequent somatic deletion of the corresponding region of the wildtype allele.52 This work has highlighted a limitation of using BAT26 alone for the detection of MMR deficiency, a limitation of considerable magnitued considering that large dele-

tions of MSH2 comprise 15% of all known deleterious MMR mutations of which up to one-half span exon 5. A third NCI workshop was convened in December 2002 to further develop guidelines for the clinical diagnosis of HNPCC and MSI testing. Results of this workshop advocated the continued use of the original NCI panel of five markers with recognition that in the future additional mononucleotide markers may be added to the panel.51

The proper differentiation between MSI-H and MSI-L/MSS tumors is important because these two tumor subsets have very different biological characteristics with different molecular etiologies. Current data indicate that the MSI-H phenotype is due almost exclusively to defective MMR. MSI-L and MSS tumors, so far, appear to share the same phenotypic characteristics with little or no involvement of defective MMR. MSI-L tumors show a genetic profile similar to sporadic MSS CRC (with typical KRAS and TP53 mutations and LOH at 5q, 17p, and 18q),but not to MSI-H CRC.53,54 In addition, several histopathologic studies demonstrate that MSI-H tumors reflect the clinical and pathological features typically associated with HNPCC (reviewed above), while MSI-L/MSS tumors are not associated with these features.49 Together, these data indicate that MSI-L/MSS CRC are fundamentally different from MSI-H CRC. This assay, therefore, provides a sensitive test for detecting the presence of an abnormality in the MMR pathway.

As noted earlier, MSI is not a specific characteristic of CRC associated with HNPCC. Large prospective studies of unselected CRC patients have determined the frequency of MSI-H in both sporadic and HNPCC-derived tumors. These studies indicated that 12% to 28% of unselected CRC display MSI.16,17 However, mutation analysis of the samples with MSI-H revealed germline MMR gene mutations in only approximately 2% of the patients. MSI also has been described in other sporadic component tumors within the HNPCC spectrum.55 These data led to an interesting paradox: although most CRC derived from individuals with a clinical diagnosis of HNPCC demonstrate MSI-H, only a minority of unselected tumors with MSI-H are from HNPCC families. These data indicate that two groups of phenotypically indistinguishable MSI-H tumors exist: those that are due to an underlying germline mutation in an MMR gene and those that are due to another mechanism. Insight into this paradox was gained, in part, by the analysis of MMR proteins in colorectal tumor specimens by immunohistochemistry (IHC). Nearly all tumors that displayed MSI-H, but without germline MMR gene mutations, demonstrated a loss of hMLH1 protein expression. Subsequent analysis of tumor tissue for hMLHl mutations indicated that somatic mutation was a relatively infrequent event. However, analysis of the methylation status of the hMLHl promoter region indicated that aberrant promoter methylation was the most likely explanation for the loss of hMLH1 expression in these tumors.56,57 Thus, it appears that MMR inactivation can account for nearly all CRC demonstrating the MSI-H phenotype.

0 0

Post a comment